regulation of medical devices by the food and drug administration

5.Regulation of Medical Devices bythe Food and Drug Administration

Hearts will never be practical until they can be made unbreakable.—The Wizard of Oz

Contents

PageIntroduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Legislative History of Device Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97The Medical Device Amendment of 1976. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Implementation of the Medical Device Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Registration of Firms and Listings of Devices . . . . . . . . . . . . . . . . . . . . . . . . . . .Premarket Notification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Classification of Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Reclassification of Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Good Manufacturing Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Performance Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Investigational Device Exemptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Premarket Approval. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Postmarketing Surveillance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Other Provisions of the Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Discussion and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Scope of Medical Device Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Regulation of Preamendments Class III Devices

and Their Postamendments Equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Regulation of Intermediate Classes of Devices . . . . . . . . . . . . . . . . . . . . . . . . .Postmarketing Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Impact of the Amendments on Medical Devices Firms . . . . . . . . . . . . . . . . . . . .Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Policy Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Scope of Medical Device Regulation, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Regulation of Preamendments Devices

and Their Postamendments Equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Regulation of Classic, or intermediate, Devices . . . . . . . . . . . . . . . . . . . . . . .Postmarketing Monitoring and Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Assistance to Small Manufacturers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

LIST OF TABLESTable No.32. Number of ’’510k” Submissions and Number Found Not

Substantially Equivalent 1976-83 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33. Classification of Preamendments Device Types by

Medical Specialty Category, February 1984.. . . . . . . . . . . . . . . . . . . . . . . . .34. Classification Status of Preamendments Device Types,

February 1984... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35. Number of Postamendments Class III Devices Found

Substantially Equivalent to Preamendments Devicesby Medical Specialty Category, 1976-83 ..,.. . . . . . . . . . . . . . . . . . . . . . . . . .

36. Investigational Device Exemptions (IDEs) for Significant Risk Devicesby Medical Specialty Category, 1977-82. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37. Approved Premarket Approval Applications (PMAAs)by Medical Specialty Category, 1977-82 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

FIGURESFigure No.1. How To Get to Market With a Medical Device . . . . . . . . . . . . . . . . . . . . . . . . .2. U.S. Patent Applications for Low-Technology Medical Devices, 1968-79 . . .3. U.S. Patent Applications for High-Technology Medical Devices, 1968-79 . . .

98102102103105107109109111112114116117118

118120120121126126127

128131132132

Page

104

106

106

107

112

113

Page

101123124

5.Regulation of Medical Devices by

the Food and Drug Administration

INTRODUCTIONThe Medical Device Amendments of 1976 (Pub-

lic Law 94-295) consolidated and expanded ex-isting Federal authority over medical devices intoa system of regulating the safety and effectivenessof medical devices in proportion to the degree ofrisk that they pose. In the past 2 years, interestin the law has grown because of problems thathave surfaced in implementing some key provi-sions and because of concerns regarding the costsof some provisions relative to the incrementalgains in safety and effectiveness.

The Subcommittee on Oversight and Investiga-tions of the House Committee on Energy andCommerce held hearings on the Food and DrugAdministration’s (FDA) implementation of thestatute in July 1982 (336) and issued an oversightreport in 1983 (338). The General Accounting Of-

fice also reviewed implementation of the Medi-cal Device Amendments and issued its report inSeptember 1983 (331). Most recently, in February1984, the Subcommittee on Health and the Envi-ronment of the House Committee on Energy andCommerce held oversight hearings on the law andits implementation (337). These hearings and in-vestigations have focused on FDA’s priorities andpace in implementing the amendments and onthose provisions of the law which, in view of theexperiences gained since the law’s enactment, havenot worked as intended.

The 1976 law, its history, its implementationby FDA, and key unresolved issues are addressedin this chapter. The chapter concludes with a pres-entation of a range of options addressed to themajor objectives of the law.

LEGISLATIVE HISTORY OF DEVICE REGULATIONMedical device regulation was first authorized

in the Federal Food, Drug, and Cosmetic Act of1938. (This act is best known for requiring pre-market notification for the safety of new drugs,a requirement that was extended to include pre-market approval of the efficacy as well as thesafety of new drugs in the Drug Amendments of1962.) The 1938 act defined medical devices as (21U.S.C. § 321 (h)):

. . . instruments, apparatus, and contrivances, in-cluding their components, parts and accessories,intended (1) for use in the diagnosis, care, miti-gation, treatment, or prevention of disease in manor other animals; or (2) to affect the structureor any function of the body of man or otheranimals.

The 1938 act authorized FDA to inspect any sitein which devices were manufactured, processed,

packed, or held (21 U.S.C. § 374). It also author-ized FDA to seize adulterated or misbranded med-ical devices; request an injunction against theirproduction, distribution, or use; or seek criminalprosecution of the responsible manufacturer ordistributor. But the agency could not take actionuntil after a device had been marketed.

In the early regulatory actions taken againstadulterated or misbranded devices, FDA was ableto use expert testimony to prove its allegations.Over time, however, FDA increasingly had to testdevices suspected of violating the law in order toremove these devices from the market (340).

As medical devices became more complex afterWorld War II, attention turned to the regulationof legitimate devices as well. But FDA could stillact only after devices were distributed and also

97

98 . Federal Policies and the Medical Devices Industry

had the burden of proving that a particular itemwas misbranded or unsafe, because devices werenot subject to premarket approval (12). In the late1960s, however, the courts ruled that certainproducts (such as nylon sutures and antibiotic-sensitive discs) that fell in the grey area betweendrugs and devices could legally be considereddrugs and subjected to premarket approval re-quirements for new drugs (12,302); subsequently,FDA regulated as “new drugs” such products assome intrauterine devices (IUDs), some contactlenses, and some in vitro diagnostic products.

Furthermore, during the late 1960s, Congressaddressed public health problems associated withradiation emissions from electronic products.Under the Radiation Control for Health and SafetyAct of 1968 (Public Law 90-602), Congress estab-lished a radiation control program to authorize theestablishment of standards for electronic products,including medical and dental radiology equipment.

From the early 1960s to 1975, six Presidentialmessages were given and 28 bills were introducedto enact medical device legislation.

A 1969 Department of Health, Education, andWelfare review of the scientific literature for in-juries associated with medical devices that wasconducted by the Cooper Committee (named afterits chairman, Theodore Cooper, then Director of

the National Heart, Lung, and Blood Institute ofthe National Institutes of Health) estimated thatover a 10-year period, 10,000 injuries were asso-ciated with medical devices, of which 731 resultedin death (339).

The vast majority of these problems were asso-ciated with three device types: artificial heartvalves, 512 deaths and 300 injuries; cardiac pace-makers, 89 deaths and 186 injuries; and intrauterinecontraceptive devices, 10 deaths and 8,000 injuries(339). As observers noted, however, there hadbeen no sensational event or public tragedy tospur more stringent regulation of medical devicessuch as the events leading to the 1962 Drug Amend-ments (165,328).

Additional examples of hazards associated withmedical devices were documented in congressionalhearings in 1973. These included prosthetic andorthopedic implants of improper materials, car-diac defibrillator with faulty electrical circuitry,incubators in which temperatures reached as highas 1450 F, plastic tracheotomy tubes with obstruc-tions, and faulty valves on emergency oxygenrespirators (339).

The developments just described eventuallyculminated in the enactment of the Medical De-vice Amendments of 1976 (Public Law 94-295).

THE MEDICAL DEVICE AMENDMENTS OF 1976The situation prior to enactment of the Medical ●

Device Amendments in 1976 was that FDA couldimpose premarket approval requirements on onlya limited number of devices that could legally be ●

considered new drugs (see above). FDA did havethe power to inspect the premises where deviceswere manufactured and distributed, but had no

●

power to require that owners of these premisesnotify FDA that they were in the device business.

to require that businesses involved with med-ical devices register their establishments andlist their devices annually,to impose regulatory requirements (standardsor premarket approval) in proportion to thedegree of risk of a device, andto impose other general controls on all de-vices to assure safety and effectiveness.

And FDA could attempt to remove mislabeled or FDA continues to have the authority granted byunsafe devices only on a case-by-case basis afterthe devices had been marketed. the 1938 act to inspect any establishment in which

devices are manufactured, processed, or packed,As a result of the 1976 Medical Device Amend- whether or not these establishments are exempt

ments, FDA currently has the authority: from registration.

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration ● 9 9

The definition of medical device was changedin the 1976 amendments to (Public Law 94-295):

. . . an instrument, apparatus, implement, ma-chine, contrivance, implant, in vitro reagent, orother similar or related article, including any com-ponent, part or accessory, which is—(1) recognized in the official National Formulary,

(2)

(3)

This

or the United States Pharmacopoeia, or anysupplement to them,intended for use in the diagnosis of disease orother conditions, or in the cure, mitigation,treatment, or prevention of disease, in man orother animals, orintended to affect the structure or any func-tion of the body of man or other animals, andwhich does not achieve any of its principalintended purposes through chemical actionwithin or on the body of manor other animalsand which is not dependent upon being metab-olized for the achievement of any of its prin-cipal intended purposes.

last clause in the definition (not achievingits principal purposes through “chemical actionwithin or on the body” and “not dependent uponbeing metabolized”) distinguishes devices fromdrugs.

Devices are to be categorized by type, on thebasis of recommendations from FDA classifica-tion panels, into three regulatory classes reflect-ing their potential risk:

● Class I—general controls,● Class II—performance standards, and● Class III—premarket approval.

Class I, general controls, encompasses devicesfor which general controls authorized by the actare sufficient to provide reasonable assurances ofsafety and effectiveness. Tongue depressors arean example. Manufacturers of Class I and all otherdevices must register their establishments and listtheir devices with FDA, notify FDA at least 90days before they intend to market the device, andconform to good manufacturing practices. Goodmanufacturing practices apply to the manufactur-ing, packing, storage, and installation of devices.

Class II, performance standards, contains de-vices for which general controls are consideredinsufficient to ensure safety and effectiveness andinformation exists to establish performance stand-ards. X-ray devices are an example.

Class III, premarket approval, applies to de-vices for which general controls are insufficientto ensure safety and efficacy, information doesnot exist to establish a performance standard, andthe device supports life, prevents health impair-ment, or presents a potentially unreasonable riskof illness or injury. Cardiac pacemakers are anexample.

Preamendments devices were to be so classifiedand placed in Class I, II, or III. Postamendmentsdevices found to be “substantially equivalent” toproducts on the market before 1976 were to beput into the same class as their preamendmentscounterparts and could be marketed immediately,although those in Class III could eventually be re-quired to demonstrate safety and effectiveness.Other postamendments devices were to be auto-matically classified into Class 111, although themanufacturer could petition FDA for reclassifica-tion into Class I or II; thus, these devices couldnot be marketed until they had completed FDApremarket approval for safety and effectiveness,

In implementing the 1976 amendments, pre-amendments Class 111 devices and their postamend-ments substantially equivalent counterparts wereto be treated differently from truly new post-amendments Class III devices. The 1976 amend-ments stipulate that manufacturers of preamend-ments Class III devices cannot be required topresent safety and effectiveness evidence until 30months after the effective date of a final classifica-tion regulation or until 90 days after publicationof a final regulation requiring submission of evi-dence on safety and effectiveness, whicheverperiod is longer (21 U.S. C. § 351(f)(2)(B)). In theinterim, preamendments Class 111 devices and theirpostamendments substantial equivalents can con-tinue to be marketed, subject only to the samegeneral controls as applied to Class I devices.

Manufacturers of any of the following devicesare required by section 510(k) of the law to notifyFDA at least 90 days prior to marketing them:

●

●

●

a device that is to be marketed for the firsttime,a device or product line that may be similarto one already marketed by another manu-facturer, ora version of an existing device in a form sig-

100 ● Federal Policies and the Medical Devices Industry

nificantly changed or sufficiently modified toaffect its safety and effectiveness.

The manufacturer’s 510k premarket notificationmust contain enough information so that FDA candetermine whether or not the device is “substan-tially equivalent” to a device already being mar-keted. To be found substantially equivalent, apostamendments device need not be identical toa preamendments device, but must not differmarkedly in materials, design, or energy source.

The legislative history reflects a congressionalintent that the term “substantially equivalent” beconstrued narrowly where necessary to assuresafety and effectiveness, but less narrowly in in-stances where differences between a postamend-ments device and a preamendments device did notrelate to safety and effectiveness (340). If FDAdetermines that a postamendments device is sub-stantially equivalent to one already in use, themanufacturer may market the device.

If FDA finds that a device is not substantiallyequivalent to one already in use before the 1976amendments, the device must go through a pre-market approval process. In this case, it is auto-matically classified into Class 111, although themanufacturer may petition FDA to reclassify itinto Class I or Class II. (Class I devices can bemarketed, subject only to the general controlssummarized earlier. Since FDA has published noperformance standards for Class II devices (seesection on “Performance Standards” below), thesedevices have been subject only to general con-trols. ) For a Class III device that is not substan-tially equivalent to a pre-1976 device, informa-tion must be provided to FDA to document itssafety and effectiveness before the device can beapproved by FDA for marketing.

In order to develop the safety and efficacy in-formation necessary for market approval of aClass III device, the sponsor of such a device mayapply to FDA for an “investigational device ex-emption” (IDE). An IDE, the parallel to the in-vestigational new drug (IND) process in drug reg-ulation, permits limited use of an unapproveddevice in controlled settings. Upon completion ofclinical investigations under the IDE, the spon-sor may submit to FDA a premarketing approvalapplication (PMAA) presenting the results of the

clinical investigations, an explanation of what thedevice consists of and how it works, manufactur-ing data that show compliance with good manu-facturing practices, and other information thatFDA may require.

If FDA approves this PMAA, the device maybe marketed. (The amendments provide an alter-native to the IDE/PMAA route to marketing ap-proval for Class III devices, called a “product de-velopment protocol, ” but this has never beenused. The major difference between the productdevelopment protocol and the IDE/PMAA proc-ess is that in the former, FDA would participatein deciding how the device is to be tested. Oncethe product development protocol is completed,the testing results would be submitted to FDA forapproval of the device for marketing (388). )

Finally, the situation for certain “transitionaldevices” (i.e., devices that were regulated as “newdrugs” before enactment of the 1976 amendments)is comparable to that for postamendments devicesthat are not substantiality equivalent to preamend-ments devices. Transitional devices are automat-ically classified into Class 111, which requirespremarket approval, but may be reclassified,subsequent to petitioning FDA, into Class I orClass II,

The current process of getting a medical deviceto market is summarized in figure 1.

The Medical Device Amendments contain otherprovisions worth noting that are applicable to allmedical devices. First, sale, distribution, or useof a device may be restricted by FDA if there can-not otherwise be reasonable assurances of itssafety and effectiveness. A device maybe bannedif it presents substantial deception or an unreason-able and substantial risk of illness or injury.

Second, manufacturers, importers, and distrib-utors of devices may be required to establish andmaintain additional records, make reports, andprovide information to FDA to assure that theirdevices are safe and effective.

Third, devices are subject to the color additiveprovisions of the Federal Food, Drug, and Cos-metic Act, but only if the color additive comesin direct contact with the body for a significantperiod of time.

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration ● 101

Figure 1.— How To Get to Market With a Medical Device

MARKET

+

general purpose articlesdevices used in research and teaching

Yes custom devices

MARKETSOURCE: U.S. Department of Health and Human Services, Food and Drug Adrninistratkm, Bureau of Medical Devices, Regulatory ~equirements for ~arketing

a Device (Washington, DC: U.S. Government Printing Office, 19S2).

102 • Federal Policies and the Medical Devices Industry

Fourth, because of concern over the impact ofthe 1976 amendments on small manufacturers, aprovision of the law stated that an office shouldbe established to provide technical assistance tosmall firms. FDA has therefore organized an Of-fice of Small Manufacturers Assistance to helpsmall firms with the regulatory requirements.

Finally, any medical device that can be mar-keted legally in the United States can be exportedlegally without further approval by the FDA.Medical devices that have not been approved foruse in the United States may also be exportedunder certain conditions. Prior FDA approval isneeded for export of devices that: 1) are in viola-tion of performance standards, 2) are subject topremarket approval, 3) are subject to limited useunder an IDE, or 4) are banned in the United

States. These four types of devices can be exportedonly if they have the approval of the country towhich the device is to be exported, and if FDAhas determined that exportation of the device isnot contrary to public health and safety (21 U.S.C.§ 381(d)(2)). &y other type of device that can-not be marketed in the United States may be ex-ported without FDA approval if the device: 1)meets the specifications of the foreign purchaser,2) does not conflict with the laws of the countryof the foreign purchaser, 3) is labeled for export,and 4) is not sold or offered for sale domestically(21 U.S.C. § 381(d)(l)). Although prior FDA ap-proval is not required, FDA can at any time re-quire the exporter of such a device to show thatthe aforementioned requirements are being met.

IMPLEMENTATION OF THE MEDICAL DEVICE LAWRegistration of Firms andListing of Devices

Federal regulations require the following busi-nesses involved with medical devices to registertheir establishments with FDA and list their de-vices annually (21 CFR pt. 807.20):

●

●

●

●

●

manufacturers and other specified processorsof devices,manufacturers of device components or ac-cessories that are ready to be used for andlabeled for a health-related purpose,initiators or developers of device specifications,repackagers and relabelers, andinitial distributors of imported devices.

Manufacturers of device components and rawmaterials who would not otherwise be requiredto register, dispensers of devices, licensed medi-cal practitioners, manufacturers of general-pur-pose articles, manufacturers of devices solely forveterinary use, and manufacturers of devicessolely for research and training are exempt fromregistration (21 CFR pt. 807.65).

The number of device establishments registeredwith FDA in 1980 was 6,073. (This number dif-fers from the number of establishments cited in

ch. 2, mainly because the FDA list includes non-manufacturing entities such as distributors. ) By1982, the number had increased to 7,636 regis-tered establishments, 6,585 domestic and 1,051foreign, listing approximately 41,500 products.More than 95 percent of the establishments hadfewer than 500 employees, and more than half hadfewer than 50 employees (143). Registration listschange significantly from year to year. In 1983,for example, 1,100 firms canceled their medicaldevice status, while 1,800 firms registered for thefirst time with FDA (206).

Two studies by FDA’s Office of Planning andEvaluation measured “baseline” conditions in or-der to track changes that may occur in the future.Some of the studies’ principal findings on deviceestablishments in 1980 were as follows (392):

●

●

Eighty-two percent of registered domesticestablishments manufactured devices, 20 per-cent imported devices, and 22 percent re-packaged devices (device establishments mayhave more than one function).Sixty-nine percent of domestic establish-ments were the sole site operated by theowner/operator, while 28 percent were sub-sidiaries, branches, or divisions.

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration • 103

●

●

●

●

●

●

Ninety-three percent of domestic owner/operators (3,948 out of 4,245 in 1980) oper-ated only one medical device establishment.Forty-two percent of the domestic establish-ments had 20 or fewer employees, while 29percent had 100 or more employees.Larger establishments were more likely thansmall establishments to: 1) produce moretypes of devices, 2) make an “exclusive” de-vice (a device made by only one or twoestablishments), 3) make a Class III device,or 4) make a “critical” device (defined belowin the section on “Good ManufacturingPractices”).Sixty-four percent of listed manufacturersmade devices in only one medical specialtyarea (as defined by FDA’s list of classifica-tion panels).Medical device establishments other thanthose making diagnostic devices averaged 4.4products each, while diagnostic device estab-lishments averaged 6.4 products each.There was little overlap between manufac-turers of medical devices and diagnostic de-vices. Establishments making dental, oph-thalmic, and radiological devices were alsohighly specialized. Therefore, there appearsto be a segmentation of the industry betweenmedical and diagnostic devices, and a furthersegmentation of the medical devices portionof the industry between establishments thatare highly specialized and those that makedevices in several areas.

The other study (391) looked at “availability”of devices, or the number of products for eachdevice type. A device type may include all prod-ucts of a particular type (e.g., cardiac pacemakers)or may include groupings of separate types ofdevices that are similar. The more products of atype, the greater the availability of products ofthat type. The analysis in this study was basedon device classifications that were establishedenough to use at the time of the analysis, ordevices from about half of the FDA classificationpanels established (see “Classification” sectionbelow). Its principal findings on availability wereas follows (391):

● On average, there were nine products pertype, i.e., each device type was made by anaverage of nine establishments.

●

●

●

●

●

Product availability was related to class ofdevice. Class I device types averaged 13.1products per type; Class II, 7.9 products; andClass III, 4.5 products.Devices with only one or two manufacturerscomprised 28 percent of all device types.Forty-one percent of Class 111, 28 percent ofClass II, and 24 percent of Class I devicetypes had only one or two manufacturers.Foreign establishments made 17 percent ofthe products examined. Eleven percent of allexclusive types had only foreign manufac-turers; 4 percent were made solely by foreignmanufacturers.Foreign products accounted for 21 percent ofClass III devices, 19 percent of Class II, and15 percent of Class I devices.More than one-third of all obstetrics-gyne-cology products and nearly two-thirds ofClass I neurological products were of foreignorigin.

Premarket Notification

In addition to listing their devices annually, de-vice establishments must notify FDA through the510k notification process (see above) when theyintend to market new devices.

Postamendments devices that are not found byFDA to be “substantially equivalent” to preamend-ments devices or to postamendments devices thathave been reclassified into Class I or II are pre-sumed to be Class III, and hence to need pre-market approval unless the device’s sponsor suc-cessfully petitions FDA to reclassify the deviceinto Class I or II. However, the overwhelming ma-jority of postamendments devices are from man-ufacturers who are marketing existing device typesfor the first time or who have devices that are mi-nor modifications of existing devices. Thus, the510k premarket notification process, together withthe FDA finding that devices are substantiallyequivalent to preamendments devices, has becomethe predominant route by which postamendmentsdevices have reached the market.

An indication of the extent to which postamend-ments devices have been regulated through the510)k notification process is reflected in the factthat, of more than 17,000 notifications receivedfor fiscal year 1977 through fiscal year 1981, only

104 ● Federal Policies and the Medical Devices Industry— .

approximately 300 were found to be not sub-stantially equivalent and therefore automaticallyplaced in Class III. For 65 of these, petitions forreclassification were received; 28 were approved,5 denied, 28 withdrawn or converted to othertypes of submissions, and 4 were still active atthe end of fiscal year 1981, Of the 28 approvals,3 were reclassified from Class III to I, and 25 fromClass III to II (143). The number of 510k submis-sions and the number of submissions found notsubstantially equivalent since 1976 are summa-rized by year in table 32.

The purpose of the 510k notification processwas to keep FDA apprised of what was going onin the industry. The concept of “substantial equiv-alence” was included in the law to address thequestion of how to treat pre- and postamendmentsdevices fairly. Two issues were involved: 1) a dou-ble standard would exist if a postamendments de-vice had to go through the premarketing approvalprocess before it could be marketed, while anidentical preamendments device would continueto be marketed; and 2) a type of monopoly wouldin effect be given to a preamendments device ifidentical pre- and postamendments devices weretreated differently.

The 510k process, together with a determina-tion of substantial equivalence, has been used ex-tensively for postamendments devices to avoid

Table 32.—Number of “510k” Submissions andNumber Found Not Substantially

Equivalent, 1976-83

NumberNumber found not

of “510k” substantiallyYear submissions equivalent

1976 (7 months). . . . . .1977 . . . . . . . . . . . . . . .1978 . . . . . . . . . . . . . . .1979 . . . . . . . . . . . . . . .1980 . . . . . . . . . . . . . . .1981 . . . . . . . . . . . . . . .1982 . . . . . . . . . . . . . . .1983 . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . .

1,3622,4272,1802,7143,3163,6523,780a

N Ab

84743447363553 2d

19,431C 365d

aEstimate.bNA indicates information not available.cExcluding 1983.‘As Of July 1983.

SOURCE: U.S. Department of Health and Human Services, Food and Drug Ad-ministration, unpublished data, Silver Spring, MD, 1983.

Class III designation and its automatic require-ment for premarket approval, or to avoid the in-volved rulemaking process necessary to reclassifysuch devices from Class III to Class I or II.

Use of the substantial equivalence clause to per-mit the marketing of devices without premarketapproval has been encouraged by FDA’s regula-tions and practices. First, FDA’s initial proposedregulations that would have required submissionof a 510k notice if modifications could affectsafety and effectiveness were changed. In the pro-posed regulations, if FDA determined that modi-fications could affect safety and effectiveness,there would be no finding of substantial equiva-lence, and an evaluation of the difference wouldhave been made in a PMAA or in a reclassifica-tion petition from automatic Class III designation(53). In the final regulation, however, FDA changedthe wording to “changes that could significantlyaffect the safety or effectiveness of the device”(emphasis added) (21 CFR pt. 807.81(a)(3)(i)).

Second, FDA allows manufacturers to traceback through a chain of substantially equivalentpostamendments devices to a device on the mar-ket before the amendments were enacted. For ex-ample, a 1982 device may be approved as substan-tially equivalent to a 1981 device, which wasapproved as equivalent to a 1979 device, and soon eventually back to a preamendments device.This practice has been labeled “piggybacking” or,alternatively, “equivalence creep” (53,331).

Third, the amount of data required to showsubstantial equivalence varies widely, dependingon the device. All devices that have been deter-mined not to be substantially equivalent andwhich thus must go through premarket approvalare reviewed centrally, but there is no such cen-tral review for devices that have been found tobe substantially equivalent (47).

Another issue relating to the 510k notificationprocess is whether or not notification require-ments should be applied to Class I devices. In Sep-tember 1982, the Scientific Apparatus MakersAssociation petitioned FDA to drop 510k notifica-tion requirements for Class I device types and tosimplify reporting requirements for Class 11 andIII (82). The petition claimed that Class I deviceswould still be subject to the registration require-

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration • 105— — —— —... — —

ments and to surveillance under the good manu-facturing practices regulations.

Furthermore, notification of intent to marketClass II devices for which no standards exist couldbe simplified, and additional information couldbe required only for Class III devices and for Class11 devices that have performance standards. Thepetition claimed that these changes would still pro-vide reasonable assurances against new devicesbeing marketed without a change in classificationor without premarket approval.

FDA subsequently denied the petition, tellingthe Scientific Apparatus Makers Association thatthe legislative intent was to make decisions on thebasis of generic types of devices, not whether ornot devices were in Class I (80). In addition, FDAwas already exempting some device types from510k notification requirements. For example, inits final rule on classifying General Hospital andPersonal Use devices, FDA exempted 30 generictypes of Class I devices from notification require-ments. They included medical absorbent fibersand specimen containers, if the devices are notlabeled or otherwise represented as sterile (306).

Classification of Devices

By the beginning of 1984, FDA had completedclassification of preamendments device types inonly 11 of its 19 medical specialty sections andhad issued proposed classifications for the other8 sections (see table 33). Final and proposed clas-sifications as of February had placed 460 devicetypes in Class I, 1,086 in Class II, 138 in ClassIII, and, depending on the particular product oruse of the product of a specified device type, 27in Class I or II, 13 in Class II or III, and one inClass I, II, or 111 (see table 34).

A “device type” may include all products of aparticular type (see discussion of availability inthe section above, “Registration of Firms andListing of Devices”) or may include groupings ofseparate types of devices that are similar. Thus,for example, the device type “obstetrics-gyne-cology specialized manual instruments” was formedby merging 18 separate instruments such as um-bilical clamps, gynecological surgical forceps, anduterine sounds (391).

Documentation of safety and effectiveness forpreamendments Class III devices was not imme-diately required but eventually has to be sub-mitted for marketing to continue. As previouslyexplained, the 1976 amendments provided a graceperiod of 30 months before such requirementscould be imposed, but the grace period does notbegin until final classification is made. Therefore,for example, the earliest date that FDA could callfor evidence of safety and effectiveness of ClassIII devices in the eight medical specialty sectionsfor which final classifications had not been madeat the beginning of 1984, even if they were finallyclassified early in the year, would be in 1986. Forthe 11 medical specialties with final classifications,the grace period had ended for 6 by 1984 (see table34).

Tables 33 and 34 show the number of Class IIIdevice types for which the 30-month grace periodapplies. An indication of the number of deviceproducts that are involved can be gleaned fromthe number of postamendments Class III devicesfound to be substantially equivalent to preamend-ments Class III devices. The number of such prod-ucts is summarized in table 35 by medical specialtyand year of notification. From the table, it canbe seen that, in addition to Class III products onthe market prior to the 1976 amendments, therewere over 1,000 postamendments Class 111 prod-ucts in use by 1983 through a finding of substan-tial equivalence. Nearly two-thirds of these prod-ucts were cardiovascular devices.

On May 5, 1982, the Health Research Grouppetitioned FDA to issue regulations requiring de-vice manufacturers to submit PMAAs for pre-amendments Class III neurological devices (252).These devices had been classified in final regula-tions effective October 4, 1979, and the HealthResearch Group had petitioned FDA shortly afterthe 30-month grace period had ended. FDA’s re-sponse was that the 30-month time period estab-lished only the earliest date FDA could act (85).

The Health Research Group subsequently wroteto Rep. John Dingell (D-Mich. ), Chair of theHouse Committee on Energy and Commerce andalso Chair of its Subcommittee on Oversight andInvestigations, asking that oversight hearings beheld and that there be consideration of an amend-

25-406 0 - 84 - 8


Table 33.-Classification of Preamendments Device Types by Medical Specialty Category, February 1984

ClassI or II or I, II,

Medical specialty category Proposed Final I II Ill II Ill or Ill Total

Neurology. . . . . . . . . . . . . . . . . . . . . . . . . . . .Cardiovascular . . . . . . . . . . . . . . . . . . . . . . . .Obstetrics-gynecology . . . . . . . . . . . . . . . . .Hematology . . . . . . . . . . . . . . . . . . . . . . . . .

}(combined) . . . . . . . . . . . . . .

PathologyGeneral hospital and personal use.............Anesthesiology . . . . . . . . . . . . . . . . . . . . . . .Immunology

)} (combined) .......................

Microbiology . . . . . . . . . . . . . . . . . . . . . . . .Physical medicine . . . . . . . . . . . . . . . . . . . . .Gastroenterology-urology . . . . . . . . . . . . . . .

Subtotal . . . . . . . . . . . . . . . . . . . . . . . . . . .Dental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .General and plastic surgery . . . . . . . . . . . . .Ear, nose, and throat . . . . . . . . . . . . . . . . . . .Ophthalmology . . . . . . . . . . . . . . . . . . . . . . .Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . .Clinical chemistry

}(combined)....................

Clinical toxicology . . . . . . . . . . . . . . . . . . .Orthopedics . . . . . . . . . . . . . . . . . . . . . . . . . .

Subtotal . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11128/784/9/794/3/799/11/79

9/11/798/2417911/2/794122/80

4122/808128/791/23/81

12/30/801/191821122/821126/8211291822112/82

2/12/82712182

9/4/792/5/802/261809/12/80

9/12/8010/21/807/16/821119/82

11191821112318311123183

——————

——

2435

42

5121

93

329

6610848

60

39105

64

4232

28049231045

7

31

15

56412225475164

175

38

180 522

460 1,086

11 024 116 0

6 0

2 27 0

5 0

2 09 4

82 713 15 0

10 04 190 0

0 0

24 0

5 6 2 0— —138 27

0 02 00 0

1 0

0 01 0

0 0

5 02 0

1 1 00 01 00 00 01 1

0 0

0 0

2 1—1 3 1

10113869

109

94134

162

8156

9441855467

11973

206

77

781

1,725

SOURCE: Federa/ Register publications of specified dates

Table 34.—Classification Status of PreamendmentsDevice Types, February 1984

Class

I or Il or I II,Status I II Ill II Ill or lll TotalFinal . . . . . . . . 280 5 6 4 8 2 7 1 1 0 9 4 4Proposed . . . . . 180 522 56 20 2 1 781— —

Total . . . . . . . 460 1,086 138 27 13 1 1,725

SOURCE: Seetable33.

ment to the device amendments that would clearlyestablish a definite time for submission of datafor preamendments and substantially equivalentClass III devices (83). The Health Research Grouppetitioned FDA again in March 1983, this timefor preamendments Class III obstetrics-gynecol-ogy devices and their substantial equivalents,pointing out that the 30-month grace period hadended on August 31, 1982 (253).

In September 1983, FDA issued its first ’’Noticeof Intent” to initiate proceedings requiring ap-proval for continued marketing of preamend-

merits Class III devices and their postamendmentssubstantial equivalents in the five medical spe-cialty categories for which the 30-month graceperiod had expired. FDA identified the followingdevices in these five medical specialty categoriesas being the first device types for which safety andeffectiveness evidence would be required (321).

● Hematology and Pathology (combined)1. Automated differential cell counter2. Automated heparin analyzer3. Automated blood cell separator

● Cardiovascular1. Implantable pacemaker pulse2. Pacemaker programmer3. Replacement heart valve

● General hospital and personal use1. Infant radiant warmer

● Neurology1. Implanted cerebella stimulator2. Implanted diaphragmatic/phrenic nerve

stimulator3. Implanted intracerebral/subcortical stim-

ulator for pain relief


Table 35.—Number of Postamendments Class Ill Devices Found Substantially Equivalent toPreamendments Devices by Medical Specialty Category, 1976-83

1976Medical specialty category (7 mos.) 1977 1978 1979 1980 1981 1982 (as of 7/83) Total

Anesthesiology a. . . . . . . . . . . . . . . . . . . . . . . . . . . 1 3 5 3 9 3 1 2 27Cardiovascular . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 71 89 121 114 143 94 51 716Clinical chemistry . . . . . . . . . . . . . . . . . . . . . . . . . 0 0 0 0 1 2 0 0 3Clinical toxicology. . . . . . . . . . . . . . . . . . . . . . . . . 0 0 0 0 0 1 0 0 1Dental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 6 5 5 4 6 11 3 42Ear, nose, and throat. . . . . . . . . . . . . . . . . . . . . . . o 3 3 6 5 0 3 0 20Gastroenterology-urology a . . . . . . . . . . . . . . . . . . 0 4 2 3 8 5 8 10 40General and plastic surgery . . . . . . . . . . . . . . . . . 6 5 2 6 5 14 2 5 45General hospital and personal usea . . . . . . . . . . 0 1 0 2 2 2 2 1 10Hematology a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . o 9 2 8 4 0 3 1 27lmmunology a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 4 1 3 3 2 1 16Microbiology a . . . . . . . . . . . . . . . . . . . . . . . . . . . o 5 4 8 12 10 3 1 43Neurology a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 6 2 3 8 5 2 0 31Pathology a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . o 0 0 0 0 0 2 0 2Obstetrics-gynecology a . . . . . . . . . . . . . . . . . . . . 2 2 3 1 3 2 0 2 15Ophthalmology . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3 7 5 3 3 9 6 38Orthopedics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . o 1 0 1 2 2 0 2 8Physical medicinea . . . . . . . . . . . . . . . . . . . . . . . . o 1 1 1 0 0 0 0 3Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . o 0 0 1 0 1 3 0 5

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .— — — —

52 124 126 175 183 202 145 85 1,092

%Iassification completed (as of the endof19S3) (see table 33).

SOURCE: U.S. Department of Health and Human Sewices, Food and Drug Administration, unpublished data, Silver Spring, MD, 19S3.

● Obstetrics-gynecologyl. Transabdominal amnioscope (fetoscope)

and accessories2. Contraceptive uterine device (IUD) and in-

troducer3. Contraceptive tubal occlusion device (TOD)

and introducer

The Federal Register notice also announced thatFDA was proposing a rule to require the filing ofa PMAA for one of these devices, the implantedcerebellar stimulator. Four months after the an-nouncement, no PMAAs had been submitted,probably because of difficulty in providing datathat supported the stimulator’s safety and effec-tiveness (86). If IDEs are obtained, however, theimplanted cerebellar stimulator may continue tobe used for the limited purpose of obtaining safetyand effectiveness data from clinical trials (321).

Reclassification of Devices

As explained earlier, the sponsors of postamend-merits devices that are not substantially equivalentto preamendments devices and are automaticallyput in Class III may petition FDA for reclassifica-tion into Class I or II. The major reclassificationissue has not been with these devices, however,

but with one of the transitional devices—contactlenses,

Under the 1976 Medical Device Amendments,transitional devices (products that had previouslybeen regulated as ’’newdrugs”) were automaticallyclassified in Class III and made subject to pre-market approval requirements, although the man-ufacturers could petition FDA for reclassification.All contact lenses made of polymers other thanpolymethyl-methacrylate (hard lenses) had beenpreviously declared to be “new drugs” and placedin Class 111 when the 1976 amendments wereenacted. Subsequently, some manufacturers didthe testing required to meet the premarket ap-proval requirements.

In March 1981, the Contact Lens ManufacturersAssociation (CLMA), representing predominantlysmall contact lens manufacturers, petitioned FDAto reclassify from Class 111 to 11 contact lenses con-sisting principally of rigid plastic materials. CLMA’scontention was that these lenses were safe and ef-fective enough to be placed in Class II, thus mak-ing further testing unnecessary.

FDA subsequently concluded that CLMA’s peti-tion did not meet all of the requirements of theregulations (21 CFR pt. 860.123). The agency also


1932 Early 1950s

Mid 1970s

Photo credits: Alan R. Kahn

This picture shows three steps in the evolution of the cardiac pacemaker, from a device carried on the patient’s back,to an external device with internal leads, to a fully implantable pacemaker Implantable cardiac pacemakers are regulated

as Class Ill devices

determined that the objective of CLMA’s petition In December 1983, FDA withdrew the proposedwas meritorious, however, and in November rule on rigid gas-permeable lenses on the basis of1982, proposed to reclassify both daily-wear soft the fact that its review found insufficient publicly

contact lenses and daily-wear rigid gas-permeable available, valid scientific evidence to show thatcontact lenses from Class III to Class I (rather than the device was safe and effective (323). The in-to Class II) (313). formation had to be based on “valid scientific


evidence” (21 CFR pt. 860.7(c)), and that evidencehad to be publicly available because the 1976amendments prohibit the use of trade secrets, con-fidential commercial information, or detailed in-formation on safety and effectiveness containedin the premarket approval application of manu-facturers who have succeeded in obtaining ap-proval for their devices.

Following its decision not to down-classify rigidgas-permeable contact lenses, however, FDAdecided to review its contact lens guidelines forIDEs and PMAAs to determine under what con-ditions some parts of the guidelines could beavoided, thereby simplifying the premarket ap-proval process (207).

Good Manufacturing Practices

Good manufacturing practices regulations,which apply to the manufacturing, packing, stor-age, and installation of devices, are one of the im-portant ways in which Class I devices were to beregulated. They also apply to Class II and IIIdevices.

The good manufacturing practices regulationsimplemented by FDA for device manufacturersdistinguish between“critical” and “noncritical”devices (21 CFR pt. 820):

“Critical device” means a device that is intendedfor surgical implant into the body or to supportor sustain life and whose failure to perform whenproperly used in accordance with instructions foruse provided in the labeling can be reasonably ex-pected to result in a significant injury to the user.

“Noncritical device” means any finished deviceother than a critical device.

Most critical devices are in Class III, but not allClass III devices are critical.

The good manufacturing practices regulationsrequire that the manufacturer keep a device mas-ter record containing the device’s specifications,production processes, and quality assurance pro-cedures; a historical record of the device indicatingcontrol numbers and dates of manufacture anddistribution; and complaint files regarding the de-vice’s performance. For critical devices, the man-ufacturer must have more detailed monitoring ofproduction and distribution and must maintain

individual control numbers and device master rec-ords. Additional compliance programs are citedspecifically for manufacturers of cardiac pacemakersand sterile devices (389). FDA has exempted man-ufacturers of some noncritical devices in Class I(e.g., specimen containers) from most of the record-keeping requirements of the good manufacturingpractices regulations.

In a review of reports of good manufacturingpractice inspections conducted primarily on Class11 and III device manufacturers from January 1979through December 1981, out of 3,811 good man-ufacturing practices inspections, 62 regulatory ac-tions were taken. FDA concluded that the com-pliance rate for larger firms tended to be somewhatbetter than for smaller firms, but overall com-pliance by the industry was good, and there wasa reasonable level of compliance for smaller firms(143).

Performance Standards

Proposed and final classifications as of early1984 had placed nearly 1,100 of the more than1,700 device types in Class 11 (see tables 33 and34, above). Yet no mandatory performance stand-ards have been issued by FDA for any Class IIdevice types.

Class 11 has become a de facto catchall regula-tory category, intermediate between the minimumregulatory requirements imposed by Class I gen-eral controls and the full premarket approvalprocess associated with Class 111 devices. Opera-tionally, however, because no performance stand-ards have been issued for Class II device types,Class II devices have been regulated as thoughthey were Class I devices.

FDA has approached further regulation of ClassII device types in several ways. First, in 1982, FDAproposed that the following steps could be con-sidered before promulgation of a mandatory per-formance standard (387):

●

●

●

●

request that manufacturers voluntarily solvedevice problems,publicize particular device problems,publish educational and technical informa-tion directed at device use,participate in developing a voluntarystandard,


● make use of other general controls such asthose for adulteration and misbranding, and

● develop guidelines.

Second, in mid-1982, the Administration sub-mitted to Congress a proposal to repeal the pres-ent statutory procedures for developing and estab-lishing performance standards for medical devicesby substituting a simpler notice-and-commentrulemaking procedure under the AdministrativeProcedures Act. The device amendments requirea five-step process: 1) initiate by Federal Registernotice a proceeding for a performance standard,which provides the opportunity for manufacturersto request a change in classification, denial of re-quests for reclassification, or initiation of reclas-sification by Federal Register notice; 2) invite per-sons by Federal Register notice to submit an ex-isting standard as a proposed performance stand-ard or an offer to develop such a standard; 3) ac-cept or reject such offers or proceed to developsuch standards; 4) publish a notice of proposedrulemaking; and 5) promulgate a performancestandard (21 CFR pt. 861.20).

The proposal the Administration submitted toCongress would have eliminated the second andthird steps. Rep. Henry A. Waxman (D-Calif. ),Chair of the Subcommittee on Health and theEnvironment of the House Committee on Energyand Commerce, agreed to sponsor the bill butadded a section requiring manufacturers to notifyFDA if they learn of device defects that presentunreasonable risks of substantial harm (see subse-quent section on “Postmarketing Surveillance” fora related discussion). H.R. 7052, the Medical De-vice Amendments of 1982, was introduced byRep. Waxman on August 19, 1982, but was notacted on. Similar legislation, including discre-tionary authority to apply performance standards,was reported to be under consideration at the De-partment of Health and Human Services/FDA atthe beginning of 1984 (87).

Third, in mid-1983, FDA finally identified 11priority Class II devices, announced its intent toproceed with development of performance stand-ards, and started the five-step process (see above)by providing the opportunity to submit a requestfor a change in the classification of the first ofthese 11 devices, the continuous ventilator (320).

Do the 1976 Medical Device Amendments infact require the use of performance standards?Two sections of the amendments seem in conflicton this point. Section 514(a)(l) of the act statesthat: “The Secretary may by regulation . . . es-tablish a performance standard for a Class II de-vice” (emphasis added). But the act’s definitionof a Class 11 device is: “A device which cannotbe classified as a Class I device because the [ClassI] controls . . . by themselves are insufficient toprovide reasonable assurance of the safety and ef-fectiveness of the device, for which there is suffi-cient information to establish a performance stand-ard to provide such assurance, and for which itis therefore necessary to establish for the devicea performance standard . . . to provide reason-able assurance of its safety and effectiveness” (em-phasis added) (§ 513(a)(l)(B)).

What if there is insufficient information toestablish a performance standard? That conditionin itself does not require Class III designation. Adevice is a Class III device if it “cannot be classifiedas a Class II device because insufficient informa-tion exists for the establishment of a performancestandard . . . and (it) is purported or representedto be for a use in supporting or sustaining humanlife or for a use which is of substantial importancein preventing impairment of human health, orpresents a potential unreasonable risk of illnessor injury” (emphasis added) (§ 513(a)(l)(0).FDA, in classifying a device into Class II, has hadto conclude that sufficient information to developperformance standards in fact exists (see the def-inition of Class II, above). Yet the fact that nomandatory performance standards have beenissued casts doubt on this conclusion.

Moreover, FDA has chosen Class II instead ofClass III designation even in some cases wherea device was of an implantable type. This is il-lustrated by proposed classifications for Generaland Plastic Surgery devices, where seven im-plantable device types (including artificial chins,ears, and noses) were proposed for Class II in-stead of Class 111 designation (311). The HealthResearch Group, commenting on these proposedclassifications, stated that implantable devicesshould be in Class III (253).


Investigational Device Exemptions

An IDE permits limited use of an unapprovedClass III medical device in controlled settings forthe purpose of collecting data on safety and ef-fectiveness. This information can subsequently beused in support of a PMAA.

The regulations that FDA has implemented onIDEs make a distinction, which is not expresslystated in the law, between “significant risk” and“nonsignificant risk” devices. A “significant riskdevice” is an investigational device that (21 CFRpt. 812.3(m)):

(1)

(2)

(3)

(4)

is intended as an implant and presents a po-tential for serious risk to the health, safety,or welfare of a subject;is purported or represented to be for a use insupporting or sustaining human life and presentsa potential for serious risk to the health,safety, or welfare of a subject;is for a use of substantial importance in diag-nosing, curing, mitigating, or treating disease,or otherwise preventing impairment of humanhealth and presents a potential for serious riskto the health, safety, or welfare of a subject;orotherwise presents a potential for serious riskto the health, safety, or welfare of a subject.

Sponsors of investigations of significant riskdevices must obtain approval by an institutionalreview board, if one exists, and must also applyfor an IDE from FDA. Investigations may notbegin until FDA approval is granted. The deter-mination of whether a device is a significant riskdevice is initially made by the sponsor. The in-stitutional review board reviewing the investiga-tional plan also makes this determination and hasthe authority to approve, require modifications,or disapprove the investigational plan. If the re-view board disagrees with a sponsor’s conclusionthat a device is a nonsignificant risk device, thesponsor has to notify FDA and apply for an IDE.

If no institutional review board exists or if FDAfinds the institutional review board’s review in-adequate, the sponsor may submit an applicationfor an IDE directly to FDA (21 CFR pt. 812.62).FDA will then decide whether an IDE is needed.The sponsor of a nonsignificant risk device neednot apply for an IDE but must obtain approval

to test the device from the institutional reviewboard of the institution where testing will occurand must meet certain reporting, recordkeeping,and monitoring requirements.

The IDE not only allows device sponsors to testthe Class III device before approval is obtainedfor marketing, but also is a method of keepingFDA apprised of the existence of clinical testing.For nonsignificant risk devices, FDA need not ac-tually be informed of the specifics of testing, andthese devices are considered to have approved ap-plications for IDEs as long as the institutional re-view board has approved the testing and certainrecordkeeping and other requirements are met (21CFR pt. 812.2(b)).

(At a December 1983 meeting of the Food andDrug Law Institute, an FDA official unofficiallyraised the idea of a written notification to FDAof the existence of a nonsignificant risk investiga-tion in addition to the normal nonsignificant riskIDE procedures. The principal purpose was to in-form FDA of the existence of clinical testing toensure that a reasonable amount of safety and ef-fectiveness information was gathered in prepara-tion for premarket approval, and to prevent man-ufacturers from profiting on unapproved devices(81).)

In a few instances, FDA guidelines have estab-lished requirements concerning the numbers of pa-tients required in a clinical study and the lengthof time they need to be followed. For example,in December 1983 FDA advised manufacturers ofYAG (yttrium aluminum garnet) lasers, a Class111 device which is used in cataract surgery, thata reasonable study population was 500 patientsstudied for 6 months and that the sponsors shouldnot add to the study without FDA approval (81).

The number of significant risk IDEs that havebeen issued from 1977 to 1982 is summarized intable 36 by medical specialty category. The num-bers in that table reflect the changing status of theIDE regulations. Until 1978, FDA required IDEapplications solely for studies of certain Class IIIdevices that had been previously regulated as newdrugs (i.e., “transitional devices”). In February1978, the IDE regulations for intraocular lensesbecame effective (21 CFR pt. 813), and IDE ap-


Table 36.—lnvestigational Device Exemptions (IDEs) for Significant Risk Devices byMedical Specialty Category, 1977-82

Medical specialty category 1977 1978a 1979 1980b 1981 1982

Anesthesiology . . . . . . . . . . . . . . . . . . . . . . . . . .Cardiovascular . . . . . . . . . . . . . . . . . . . . . . . . . . . .Clinical chemistry . . . . . . . . . . . . . . . . . . . . . . . . .Clinical toxicology ., . . . . . . . . . . . . . . . . . . . . . . .Dental ............ . . . . . . . . . . . . . . . . . . .Ear, nose, and throat . . . . . . . . . . . . . . . . . . . . . . .Gastroenterology-urology . . . . . . . . . . . . . . . . . . .General and plastic surgery . . . . . . . . . . . . . . . . .General hospital and personal use . . . . . . . . . . .Hematology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Immunology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Microbiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Neurology . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . .Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Obstetrics-gynecology. . . . . . . . . . . . . . . . . . . . . .Ophthalmology . . . . . . . . . . . . . . . . . . . . . . . . . . .Orthopedics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Physical medicine . . . . . . . . . . . . . . . . . . . . . . . . .Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

000000042000000

0o0

7

030000010010000

48000

53

o 02 212 00 01 40 11 102 50 40 00 20 13 20 00 21

21 240 41 00 6

33 105

8 1139 320 10 00 21 4

28 217 124 40 02 00 04 90 0

53 3434 3412 11

2 09 1

203 176

Total

1997

3076

603114

051

180

108C

16227

316

577alntraocular lens regulation final in February 1978blDEregula~onf inal in January 1980.cAlmost exclusively cervical cap studies.

SOURCE: U.S. Department of Health and Human Services, Food and Drug Adminmtration unpubhshed data, Sflver Spring, MD, 1983

placations for intraocular lenses began to be re-ceived. In January 1980, IDE regulations appli-cable to other types of devices were made final(21 CFR pt. 812).

The relationship between requests for approvaland FDA’s finding that IDEs were in fact neededis indicated by the fact that nearly 60 percent ofapproximately 400 requests for approval of IDEsfor significant risk devices that FDA received be-tween July 1981 and July 1982 were approved withor without additional conditions within 30 days.The remainder were disapproved, subject toad-ditional justification, withdrawn by the sponsor,or returned to the sponsor with the finding thatan IDE was not necessary (143).

Premarket Approval

In 1980, FDA developed guidelines for the sub-mission of PMAAs and also published proposedregulations on premarket approval requirements(308). However, the regulations had not beenfinalized by early 1984. Under the guidelines cur-rently inuse, when a PMAA is approved, the ap-proval letter states that information on adversereactions and device defects must be reportedwithin 10 days. And according to an FDA pre-scription device regulation (21 CFR pt. 801.109)

predating the 1976 amendments, certain devicesmay be sold or distributed only by or on the or-der of licensed practitioners. FDA has used theserestrictions as a condition of approval for certaindevices.

As indicated earlier, the only types of devicesthat have had to go through the full premarketapproval process so far are: 1) postamendmentsdevices that are not substantially equivalent topreamendments devices, and 2) ’’transitional de-vices” that have not been reclassified as Class Ior II and postamendments devices substantiallyequivalent to them. Preamendments Class IIIdevices and their postamendments equivalents willeventually have to go through a similar approvalprocess. In September 1983, FDA identified thefirst 13 preamendments Class III device types forwhich evidence of safety and effectiveness willsoon be required if continued marketing is to beallowed (see section on “Classification of De-vices, ” above) (321).

The number of postamendments Class III de-vices that have successfully passed through thefull premarket approval process and the numberof transitional devices that have received pre-market approval from 1977 to 1982 are summar-ized in table 37 by medical specialty category.


Table 37.—Approved Premarket Approval Applications (PMAAs) by MedicalSpecialty Category, 1977-82

Medical specialty category 1977 1978 1979 1980 1981 1982 Total—Approved PMAAs:Anesthesiology . . . . . . . . . . . . . . . . . . . . . . . . . . .Cardiovascular . . . . . . . . . . . . . . . . . . . . . . . . . . . .Clinical chemistry . . . . . . . . . . . . . . . . . . . . . . . . .Clinical toxicology . . . . . . . . . . . . . . . . . . . . . . . . .Dental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Ear, nose, and throat . . . . . . . . . . . . . . . . . . . . . . .Gastroenterology–urology . . . . . . . . . . . . . . . . . . .General and plastic surgery . . . . . . . . . . . . . . . . .General hospital and personal use . . . . . . . . . . .Hematology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Immunology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Microbiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Neurology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Obstetrics-gynecology . . . . . . . . . . . . . . . . . . . . .Ophthalmology . . . . . . . . . . . . . . . . . . . . . . . . . . .Orthopedics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Physical medicine . . . . . . . . . . . . . . . . . . . . . . . .Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Approved PMAAs for new devices:Anesthesiology . . . . . . . . . . . . . . . . . . . . . . . . . . .Cardiovascular . . . . . . . . . . . . . . . . . . . . . . . . . . . .Clinical chemistry . . . . . . . . . . . . . . . . . . . . . . . . .Clinical toxicology . . . . . . . . . . . . . . . . . . . . . . . . .Dental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Ear, nose, and throat . . . . . . . . . . . . . . . . . . . . . . .Gastroenterology-urology. . . . . . . . . . . . . . . . . . .General and plastic surgery . . . . . . . . . . . . . . . . .General hospital and personal use . . . . . . . . . . .Hematology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Immunology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Microbiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Neurology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Obstetrics-gynecology . . . . . . . . . . . . . . . . . . . . .Ophthalmology . . . . . . . . . . . . . . . . . . . . . . . . . . .Orthopedics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Physical medicine . . . . . . . . . . . . . . . . . . . . . . . . .Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Approved PMAAs for transitional devices:Anesthesiology . . . . . . . . . . . . . . . . . . . . . . . . . . .Cardiovascular . . . . . . . . . . . . . . . . . . . . . . . . . . . .Clinical chemistry . . . . . . . . . . . . . . . . . . . . . . . . .Clinical toxicology . . . . . . . . . . . . . . . . . . . . . . . . .Dental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Ear, nose, and throat . . . . . . . . . . . . . . . . . . . . . . .Gastroenterology-urology . . . . . . . . . . . . . . . . . . .General and plastic surgery . . . . . . . . . . . . . . . . .General hospital and personal use . . . . . . . . . . .Hematology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Immunology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Microbiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Neurology ....., . . . . . . . . . . . . . . . . . . . . . . . . . .Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Obstetrics-gynecology . . . . . . . . . . . . . . . . . . . . .

o1000000000000000007

01000000000000000007

000000000000000

00000000000000000000

00000000000000000000

o00000000000000

0400101000050006200

19

o300101000050000200

12

o10000000000000

120000041034000

12100

28

1200000010040000100

9

o00000040030000

331010042002100

22100

40

3210000220021000000

13

o100

0o20000000

450000222001200

20200

40

4500002120012000100

18

o00000010000000

815

10203

10503

12300

60600

128

813

101033500

123000

4o

53

o20010070030000


Table 37.—continued

Medical specialty category 1977 1978 1979 1980 1981 1982 Total

Ophthalmology . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 0 6 12 22 20 60Orthopedics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 0 0 0 2Physical medicine . . . . . . . . . . . . . . . . . . . . . . . . . 0 0 0 0 1 1 2Radiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 0 0 0 0 0 0

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 0 7 19 27 22 75SOURCE: ~.~. ~:epatment of Health and Human Setvices, Food and Drug Administration, unpublished data, Silver Spring,

Approved transitional devices are heavily skewedtoward ophthalmic products, which are almostexclusively contact lenses and contact lens clean-ing solutions and, beginning in December 1981,intraocular lenses. Other examples of transitionaldevices include cardiovascular grafts, bone ce-ment, absorbable sutures, and specific types ofimmunological tests.

“New” postamendments devices are also con-centrated in a few medical specialties, but not tothe extent that ophthalmic devices have domi-nated approved transitional devices. Among thesenew devices are cardiac valves, heart pacemakersand accessories, cardiovascular catheters, life-support monitoring systems, implantable infusionpumps, artificial hips, and antibody tests for in-fectious agents.

From 1977 through the end of 1982, 128 ClassIII products (53 “new” devices and 75 “transi-tional” devices) had gained premarket approval.During the same period, 1,007 Class III productswere approved for marketing through the find-ing that they were substantially equivalent topreamendments Class 111 device types (comparetables 35 and 36). As previously noted, evidenceof safety and effectiveness has not yet been re-quired for these postamendments Class III prod-ucts found substantially equivalent to preamend-ments devices. Many of these applications forpostamendments Class 111 devices are for modifi-cations of devices which were already commer-cially available.

Postmarketing Surveillance

There are a number of existing and potentialmethods to monitor hazards associated with theuse of devices that have been marketed. FDA

maintains a Device Experience Network (DEN)that receives voluntary reports on device hazards;can require repair, refund, or replacement of de-vices for hazards or defects; and requires thatmanufacturers keep records of complaints as partof the good manufacturing practices regulations.Two other methods have been mentioned earlier.A condition of approval for new Class III devicesapproved through the full premarket approvalprocess is that information that manufacturers re-ceive on device defects and adverse reactions hasto be reported to FDA within 10 days. And man-ufacturers, importers, and distributors of devicesmay be required to provide FDA with informa-tion to ensure that their devices are safe and ef-fective.

The major issue in postmarketing surveillanceactivities has involved the authority that the 1976amendments gave to FDA to require that infor-mation be provided to FDA to ensure that devicesalready on the market are safe and effective. Inlate 1980, FDA proposed rules for mandatory de-vice experience reporting, under which manufac-turers and distributors of medical devices wouldbe required to submit reports on devices that: 1)may have caused a death or injury, 2) may havea deficiency that could cause a death or injury orthat could give inaccurate diagnostic informationthat could result in improper treatment, or 3) arethe subject of a remedial action by the manufac-turer (307). Any death that might have beencaused by a device would have had to be reportedwithin 72 hours of the manufacturer’s or distrib-utor’s receipt of that information, with a followupreport submitted within 7 working days. Reportsalso would have had to be submitted within 7working days after receiving information of anyactual or possible device deficiency that could re-sult in a death or injury.


FDA’s rationale for a mandatory device report-ing regulation was twofold. Practitioners andusers of medical devices usually do not report de-vice experiences to FDA but instead contact themanufacturer for information and advice. Reportswould be required even if the manufacturer deter-mined that the death or injury was not due to thedevice or that there was no deficiency, becauseFDA expected that few devices would be charac-terized by their manufacturers as having deficien-cies, few reports would be submitted, and report-ing of confirmed deficiencies would be delayedif manufacturers first investigated before report-ing (307).

A year later, in late 1981, the proposed rule washeld in abeyance because of comments that therequirements were overly broad and because ofissuance in early 1981 of Executive Order 12291on “Federal Regulation, ” under which regulatoryactions are to be taken only when the potentialbenefits of the action outweigh the potential costs.FDA also announced that it would inspect com-plaint files maintained under the good manufac-turing practices regulations to determine if theycould be used as an adequate or partial substi-tute for the proposed rule (309). A pretest, phaseI of the review of good manufacturing practicescomplaint files was completed on December 31,1981, and phase II started on July 14, 1982, in-volving a review of the complaint files of 418firms.

In May 1983, FDA issued a reproposal on med-ical device reporting, under which reports wouldbe required within 15 days of receiving informa-tion that “reasonably suggests, or a person allegesand the manufacturer or importer is aware of theallegation” that one of its marketed devices “hascaused or contributed to” a death or serious in-jury or “has malfunctioned” and, if the malfunc-tion occurs, “is likely to” cause or contribute toa death or serious injury (318).

Data analysis of the phase 11 review of goodmanufacturing practices complaint files had beencompleted by early 1984, and the report was ex-pected to be available sometime in 1984. FDA hasconcluded that the good manufacturing practicescomplaint files would not be an adequate substi-tute for a mandatory device reporting regulationfor several reasons.

First, inspection of complaint files would notlead to timely reporting. Good manufacturingpractices inspections are conducted every 2 years,and FDA did not expect more frequent inspectionsin the future. Second, with over 6,000 establish-ments to inspect, there would be a problem withdeciding which and how many establishments toinspect. Third, the way in which good manufac-turing practices records are kept would lead topractical difficulties in collecting the informationfor adverse experience information. As a conse-quence, FDA expects to reissue a revised man-datory device reporting proposal in 1984, subjectto clearance by the Office of Management andBudget and other Federal agencies (257).

FDA, in its proposals for a mandatory devicereporting regulation, has stated that its voluntaryreporting system—the Device Experience Net-work, or DEN—is not an adequate substitute.DEN is not a comprehensive reporting program,and FDA does not have the resources to main-tain constant contact with all device users to en-courage reporting. Furthermore, device manufac-turers are the most knowledgeable about theirproducts and their associated risks and are in thebest position to report to FDA. But few manu-facturers report under the DEN system, and manyof the reports that they make are trade complaintsabout a competitor’s product, not reports fromthe manufacturers of the devices in question. Andin some cases, device manufacturers report deviceproblems to FDA only after a product recall orother remedial action is completed (318).

Reviews of DEN data on Class III devices andof recalls prompted by a hazard with a highlikelihood of serious injury or death resulted inthe following observations (based on informationprovided OTA by FDA for the period from 1976to mid-1983). From the DEN system: deaths al-legedly associated with devices were reportedmost frequently for pacemakers and heart valves;actual injury, reported most frequently with pace-makers, heart valves, IUDs, and to a lesser de-gree but still relatively frequently, with intraocularlenses; and potential injury, reported most fre-quently with resuscitation equipment (usuallyassociated with power failure or other electricalmalfunction), intra-aortic balloon pumps or cath-eters, pacemakers, heart valves, and intraocularlenses.


Recalls prompted by risks of serious injury ordeath were most frequent for cardiovascular de-vices, with pacemakers again comprising thelargest subgroup. Thus, the DEN system andrecalls for high risks mostly involve implantabledevices, often involve electrical problems, andoften involve cardiovascular devices.

The DEN system of voluntary reporting andproduct recall information do not provide ade-quate information on the magnitude and frequencyof device-related problems. Voluntary reportingalso includes allegations of death or injury thatmay not be associated with the device in ques-tion or may be user-related and not due to de-vice defects. FDA also cautions against using DENfor trend analysis, because reports are voluntary,use of the system has changed over time, and thenumber of reports therefore may reflect trends inDEN participation and other factors (48). How-ever, voluntary reporting does provide indicationsof the types of devices that have associated risks,and product recall information identifies deviceswith significant actual or potential risks.

Other Provisions of the Law

Restricted Devices

Section 520(e) of the Medical Device Amend-ments added a provision for “restricted devices”authorizing FDA to issue regulations imposingrestrictions on the sale, distribution, or use ofdevices. FDA was also authorized to regulateadvertising of restricted devices and to inspectmanufacturers’ records related to restricted de-vices. Prior to the amendments, the sale and dis-tribution of some devices were authorized onlythrough “prescriptions” by designated persons(e.g., physicians) (21 CFR pt. 801.109). Immedi-ately following the enactment of the law, FDApublished a notice announcing that FDA consid-ered “restricted devices” to include all “prescrip-tion devices” (303).

When FDA attempted to inspect the records forsome prescription devices, however, some man-ufacturers refused to comply, claiming that FDAhad to first issue regulations designating prescrip-tion devices as restricted devices. The U.S. DistrictCourts involved in resolving this issue ruled for

the manufacturers, and both the First Circuit andSecond Circuit of the U.S. Court of Appeals af-firmed the decisions of the lower courts (30,171).

As a consequence, FDA decided to issue a reg-ulation rather than attempt to establish throughfurther litigation its authority to inspect recordsfor restricted devices.

The proposed rule on restricted devices waspublished in October 1980 (305). However, FDAwithdrew the proposed rule in November 1981,stating as its reasons: 1) comments that the cur-rent prescription device regulation was sufficient,and 2) the February 17, 1981, Executive Order12291 on “Federal Regulation” that required Fed-eral agencies to undertake regulatory actions onlywhen the potential benefits of the action to societyoutweigh the potential costs. FDA also stated thatit would use the authority for inspection of recordsrequired by the good manufacturing practices reg-ulations, as well as the dispensing and labelingrequirements of the prescription device regula-tions, in lieu of a restricted device regulation (309).

Banned Devices

The banned device provision of the law hasbeen used once. Prosthetic hair fibers intended forimplantation into the human scalp were bannedin June 1983 (319,324).

Color Additives

FDA has not issued regulations on the color ad-ditive provisions of the amendments, but the issuehas so far been limited primarily to tinted con-tact lenses. All contact lenses that are required tohave premarket approval are also subject to thecolor additive provisions of the law. FDA initiallyapproved tinted contact lenses even though thecolor additives had not been listed for that usebefore the applications were approved (317).When FDA subsequently concluded that it hadto apply the color additive provision to tinted con-tact lenses, it decided that the least unfair methodwas to complete action on the pending PMAAsand to enforce the provision with future PMAAs(323). FDA is also developing proposed changesin the procedural regulations for color additivesto govern their use in all applicable devices (323).


Export of Devices

FDA regulations have not had a great effect onexport of medical devices, because most exporteddevices are those that are legally marketed in theUnited States and require no special FDA ap-proval.

Most devices requiring FDA approval for ex-port are devices that require but have not yet re-ceived premarket approval. The requirement thatthe importing country approve imports posedsome problems because of the possibility thatthere might not be an official who could give ap-proval. For that reason, FDA has accepted, in lieuof an express approval, a statement from theforeign government that it has no laws prohibitingimportation of the device in question. From Oc-tober 1, 1981 through March 31, 1983, 376 med-ical devices were approved for export, 13 deviceswere disapproved, and 1 previous approval wasrescinded (181).

Assistance to Small Manufacturers

Both the prevalence and absolute magnitude ofregulatory costs increase with establishment size,

DISCUSSION AND CONCLUSIONS

The principal provisions of the statute andFDA’s activities have been cataloged above. The1976 Medical Device Amendments attempted toregulate medical devices in proportion to a de-vice’s degree of risk through a number of pre- andpostmarketing controls. The amendments placedimmediate regulatory priority on significantly newdevices while providing a grace period beforesubstantial evidence of safety and effectivenesshad to be provided for preamendments Class IIIdevices and their postamendments equivalents.

Section-by-section descriptions and analyses ofthe major provisions of the Medical Device Amend-ments and their implementation by FDA wereprovided above to yield an understanding of theexperience so far with the regulation of medicaldevices. The analysis also identified specific reg-ulatory actions that have been proposed as alter-natives to the current situation. However, iden-

but the costs of regulations appear more unfavor-able to the small manufacturer when costs are con-sidered in proportion to establishment size (197).For example, in a limited study of 30 companiesmanufacturing only cardiovascular, anesthesiol-ogy, or diagnostic products, both initial and re-curring costs per employee were higher for smallplants (fewer than 100 employees) than for largerplants (100 or more employees) (17).

Small manufacturers also are more likely toneed assistance in complying with the regulatoryrequirements. FDA’s Office of Small Manufac-turers Assistance has received favorable reviewsby manufacturers. In a survey of medical devicemanufacturers, over three-quarters had heard ofthe office, about half of those who had heard ofit contacted it, and more than three-quarters ofthose contacting the office had found it helpful(197).

tifying specific regulatory areas and analyzing thecurrent approaches (and limitations) and theiralternatives are not the same as developing strat-egies (including maintaining the status quo) formedical device regulation. The relative signifi-cance of actions that could be taken in specificareas of medical device regulation is hard to de-termine and justify without relating these actionsto more specific strategies than the general rubricof meeting safety and effectiveness objectives atminimal regulatory costs.

In the following analysis, the principal issuesthat have arisen in implementing the Medical De-vice Amendments of 1976 are examined. Areasto be discussed include:

● the scope of medical device regulation,● regulation of preamendments Class III de-

vices and their postamendments equivalents,

118 ● Federal Policies and the Medical Devices Industry———.

• regulation of intermediate classes of devices,● postmarketing controls, and● impact of the amendments on medical device

firms.

Scope of Medical Device Regulation

FDA has exempted firms from certain require-ments of the 1976 Medical Device Amendments;under FDA’s IDE regulations, for example, adistinction is made between “significant” and“nonsignificant” risk devices, and sponsors of in-vestigations of nonsignificant risk devices obtainan IDE from an institutional review board ratherthan from FDA (see section on “InvestigationalDevice Exemptions, ” above). The law also ex-pressly permits FDA to exempt firms from noti-fying FDA about their intent to market selecteddevices, and FDA has done so for selected typesof Class I devices, subject to minimal recordkeep-ing requirements.

The Scientific Apparatus Makers Associationhad petitioned FDA to drop notification require-ments for Class I devices, claiming that Class Idevices would still be subject to the registrationrequirements and surveillance under the goodmanufacturing practices regulations. FDA subse-quently denied the petition on the grounds thatthe legislative intent was to make decisions on thebasis of generic types of devices, and not whetheror not devices were in a specific class (see sectionon “Premarket Notification, ” above).

Rather than being considered on the basis ofthe present statute’s legislative intent, the proposalfor dropping notification requirements for ClassI devices could be reconsidered in a reassessmentof the statute. With over 7,000 device establish-ments registered with FDA, listing approximately41,500 products representing over 1,700 devicetypes, one important question that arises is whetherthe scope of present device regulation is too broad.Not only could regulatory costs be excessive wheninformation is gathered that is not going to beused, but other activities undertaken to helpassure safety and effectiveness could be curtailedbecause of competition for funds within a limitedFDA budget.

Regulation of Preamendments Class IllDevices and Their PostamendmentsEquivalents

As of early 1984, classifications had been com-pleted for device types in 11 of the 19 medical spe-cialty categories, and proposed regulations, mostinitially issued in 1982, had been issued for thosein the remaining 8 (see table 34). As preamend-ments Class III devices and their postamendmentsequivalents cannot be required to show substan-tial evidence of their safety and effectiveness un-til at least 30 months after final classification, itwill be 1986 at the earliest before manufacturersof devices in the eight medical specialty catego-ries without final classifications can be requiredto show that their products are safe and effective.

FDA could have expedited classification ofhigh-priority device types within each medicalspecialty category instead of waiting to classifyall devices within each category. For example, theclassification process for device types that hadbeen provisionally designated Class 111 could havebeen completed first, thereby starting the clockon the 30-month grace period.

On the other hand, the medical specialty cate-gories for which FDA first issued final classifica-tions (see table 34) include the categories in whichmost of the deaths and injuries were found in areview of the literature by the Cooper Commit-tee before the amendments were enacted—i.e.,cardiovascular (heart valves, pacemakers) andobstetrics-gynecology (IUDs). Devices in thesecategories continue to be the major causes of deathor serious injury as reported in FDA’s voluntaryDEN reporting system (see section on “Postmar-keting Surveillance, ” above). Thus, the medicalspecialty categories for which FDA has completedclassification include those categories containingdevices with the highest known risks.

Related to classification of preamendments de-vices is the regulation of similar postamendmentsdevices through application of the “substantialequivalence” clause and the practice of “equiva-lence creep” or “piggybacking” whereby a post-amendments device can be found “substantiallyequivalent” to another postamendments devicethat had been previously found to be substantially

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration ● 119— . .

equivalent to an actual preamendments device.One issue is the safety and effectiveness of post-amendments Class III devices that have been per-mitted to be marketed through the “substantialequivalence” route, because the preamendmentsdevices against which they have been comparedhave yet to be required to show evidence of theirsafety and effectiveness.

After the 30-month grace period for preamend-ments devices expires, if FDA requires evidenceof safety and effectiveness for their continued mar-keting, more evidence on safety and effectivenesswill be available on the preamendments deviceswith which postamendments devices are com-pared. Each manufacturer of a Class III device,whether pre- or postamendments, would have tosubmit a PMAA if required, because FDA can-not consider the evidence of safety and effective-ness in one application when reviewing anotherdevice, even one that was previously found sub-stantially equivalent.

As discussed earlier, FDA has initiated pro-ceedings for some preamendments Class III de-vices for which the grace period has ended andwhich FDA has determined have the highest needfor evidence of safety and effectiveness (e.g., theimplanted cerebella stimulator). Criticisms of thepace at which FDA classified preamendmentsdevices, which determines when evidence of safetyand effectiveness of preamendments Class IIIdevices could be required, could have been mutedif the classification process had been speeded upfor preamendments Class III devices in all cate-gories. Final classification of preamendmentsdevices is no longer a major issue, however, be-cause classifications have already been proposedfor those medical specialties without final classif-ication (see table 34), and final classificationshould occur soon.

The remaining issues are: 1) what type of safetyand effectiveness evidence should be required forpreamendments Class 111 devices; and 2) how the“substantial equivalence” clause should be appliedby FDA. FDA, in announcing its intent to requiresafety and effectiveness evidence for those pre-amendments Class III devices it has identified ashaving high priority, indicated that it intended toask for data of the type needed for premarket ap-

proval of new postamendments Class 111 devices(321). However, for less controversial preamend-ments devices, more flexibility in the types ofevidence that have to be provided maybe appro-priate. As for the application of the “substantialequivalence” clause, other interpretations or othermethods of approving postamendments Class IIIdevices are possible (see “Policy Options” sectionbelow).

As noted above, the regulations that FDA hasissued on IDEs distinguish between “significantrisk” devices, for which sponsors have to receiveexpress approval from FDA to conduct studiesunder an IDE, and all other Class III devices, ofwhose testing FDA need not be actually informedand which are considered to have approved IDEssubject to certain conditions (see section on “In-vestigational Device Exemptions, ” above). Thisdistinction reflected express statutory authorityand a decision by FDA that risks from Class IIIdevices varied and monitoring of testing shouldreflect the degree of risk.

FDA has also indicated that IDEs will be madeavailable to manufacturers of preamendmentsClass III devices so that they can continue to mar-ket their devices if they cannot provide reason-able assurance of safety and effectiveness whenFDA requests such information. In its “Notice ofIntent to Initiate Proceedings to Require PremarketApproval of Preamendments Devices,” FDA hasstated that within 90 days of the issuance of a finalregulation, a PMAA must be filed or commercialdistribution has to cease.

But an alternative for the manufacturer is toobtain an IDE and continue distribution for thelimited purpose of obtaining safety and effective-ness data from clinical trials. In addition, undersection 515(6) of the amendments, FDA can ex-tend the grace period if it finds that “the continuedavailability of the device is necessary for the pub-lic health” (321).

The rationale for this use of the IDE is weak.Manufacturers of preamendments devices havehad years to prepare to substantiate the safety andeffectiveness of their devices, because the law waspassed in 1976, the classification process is stillnot over, and there is a 30-month minimum graceperiod from the date of final classification.


Regulation of Intermediate Classesof Devices

For several reasons, Class II designation hasprobably received the most attention. First, ClassII represents the important middle ground of thewhole regulatory approach. Second, the majorityof device types have been placed in Class II (morethan 1,000 out of over 1,700; see table 34). Andthird, no performance standards have yet beenissued.

Regardless of whether or not the 1976 statuterequires, rather than permits, the use of perform-ance standards, the fact remains that, as a prac-tical matter, there is little possibility that stand-ards can be formulated for the large number ofdevice types that have been placed in Class II. Ifperformance standards were meant to be selec-tively used, the designation of so many devicetypes as Class 11 and the resulting perception ofthe futility of such an exercise have been damag-ing to FDA’s efforts, no matter what the rationale.

At the least, the present situation points out theneed for an intermediate regulatory class, the in-appropriateness of mandatory performance stand-ards as the sole or even principal method of reg-ulation, and the need for other methods of regu-lating intermediate devices.

There are, of course, many ways of regulatingan intermediate class of devices. The principalissues here are: 1) whether a change in the stat-ute is needed before FDA can use other than per-formance standards, and 2) what types of regu-latory controls could be used.

Postmarketing Controls

Postmarketing controls on medical devices areof two types: 1) removal from the market orrestrictions on the sale, distribution, or use of des-ignated devices; and 2) postmarketing surveillanceof the clinical experiences with medical devices.

FDA can remove a device from the market byrequiring repair, refund, or replacement; by ban-ning it; or by revoking any approval to marketthe device. There are two types of restrictions onthe sale, distribution, or use of a device. The firstis a restriction to prescription sale or use, applied

when adequate labeling for lay use cannot be writ-ten or when special skills or training are required,such as diagnosing a disease or condition or pre-scribing for treatment. The second is a restrictionbased on other conditions FDA may prescribe inregulations in order to provide reasonable assur-ance of the safety and effectiveness of the device.

The restricted device regulations were with-drawn by FDA with the explanation that theprescription device regulations were adequate. Inaddition, Executive Order 12291 requires that Fed-eral agencies undertake regulatory actions onlywhen the potential benefits outweigh potentialcosts. However, in the original proposed regula-tion, FDA had stated that (305):

. . . the current determination that a device is a“prescription” device is quite subjective. Often,the determination is made by the manufacturer.

FDA therefore proposed the restricted device ruleto make these criteria more objective.

FDA also stated in its withdrawal of the pro-posed restricted device rule that it would use itsinspection authority under the good manufactur-ing practices regulations to inspect manufacturers’records on these types of devices for informationon such matters as deaths and injuries (309). Butthe Subcommittee on Oversight and Investiga-tions of the House Energy and Commerce Com-mittee observed that (338):

. . . most of the general controls . . . are gearedtoward ensuring that finished devices, when readyfor use, will be free from defects, safe and effec-tive. Restriction, on the other hand, can addressproblems with a device once it is in use. It dealswith the risks that practitioners, technicians, orothers who employ the device are doing so im-properly due to inadequate training, experience,facilities, or instructions.

These issues—use of existing sources of infor-mation on deaths and injuries, and problems aris-ing from improper use of medical devices ratherthan from improper manufacture—have also beeninvolved in the debates on the types of postmar-keting monitoring activities that should be con-ducted.

One of the expressed reasons why mandatorydevice reporting regulations have been held inabeyance was to examine whether the complaint

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration • 121— — —. — —

files which are required under the good manufac-turing practices regulations could partially orcompletely substitute for the mandatory devicereporting regulations. As described earlier, the ex-amination had been completed by early 1984, withthe conclusion that the good manufacturing prac-tices complaint files are not adequate substitutesfor mandatory reporting (257).

On the question of improper use, the GeneralAccounting Office has recommended that FDA’svoluntary DEN reporting system be revised sothat information is included on the scope andnature of device problems caused by user errorand inadequate maintenance; that the data beanalyzed to identify special problems, areas whereproblems might be concentrated, and trends; andthat the results be used to aid in developing solu-tions. FDA responded that these recommenda-tions would be taken into consideration and thatpossible actions would include implementingeducational programs or restricted use criteria(331).

Thus, FDA may eventually issue restricted de-vice regulations, subject to the current adminis-tration’s position on deemphasizing regulatory ap-proaches and its preference for voluntary initiatives.Furthermore, efforts may be made to upgradethe voluntary DEN system and disseminate thatinformation to educate users about potentialhazards.

Impact of the Amendments onMedical Devices Firms

The preceding analyses examined individualprovisions of the 1976 Medical Device Amend-ments and their implementation by FDA. A broaderissue is the impact of the amendments on the med-ical devices industry. Information available on theimpact of the law reflects regulatory implemen-tation by FDA and understanding of the law bydevice firms in the first few years following pas-sage of the statute. In evaluating this impact, itis important to keep in mind that some sectionsof the amendments have been implemented fully,some partially, and some have yet to be addressed.

Two studies that FDA conducted to establish“baseline conditions” in order to track changes

that occur in the future were summarized earlier,based primarily on 1977 data (391,392). Thesestudies showed that the “medical devices indus-try” is quite heterogeneous. There is a clear sep-aration of the industry into medical device versusdiagnostic testing firms, with little overlap be-tween manufacturers of either type of product.

The medical device portion of the industry isfurther separable into establishments that arehighly specialized and those that manufacturedevices in several areas. Sixty-four percent ofmanufacturers made devices in only one devicearea. Highly specialized areas include dental, oph-thalmic, and radiological devices.

Each device type is made by an average of ninedifferent manufacturers, but this measure of “prod-uct availability” or “concentration” in the indus-try is related to the class of the device. Class I de-vice types averaged 13.1 manufacturers per type;Class II,. 7.9 manufacturers; and Class III, 4.5manufacturers. Devices made by only one or twomanufacturers (“exclusive” devices) comprised 28percent of all device types and followed a similarpattern. Only one or two establishments weremanufacturing 41 percent of the Class III devicetypes, compared to 28 percent of Class II devicetypes and 24 percent of Class I device types.

Large establishments were more likely to make:1) more device types, 2) an “exclusive” device, 3)a Class III device, or 4) a “critical” device (definedby FDA as requiring more rigorous controls inthe manufacturing process).

These findings lead to the following observa-tions. First, the distinction between firms thatmanufacture diagnostic tests and firms that man-ufacture other medical devices probably reflectsthe “catchall” nature of the 1976 Medical DeviceAmendments, which essentially authorized Fed-eral regulation over all medical products that arenot drugs or biologics. One question is the ap-propriateness of regulating such distinctly differentproducts in a similar manner. For example, ClassIII medical devices are generally those that are im-planted or have life-support or life-sustainingfunctions, and criticisms of FDA’s application ofthe law to devices of this nature have been raisedwhen FDA has chosen not to place some of thesetypes of devices in Class 111 (253).

25-406 0 - 84 - 9


On the other hand, diagnostic tests pose fewdirect risks, but some have been placed in ClassIII because defective tests could lead to erroneoustreatment (or no treatment), which in turn couldresult in harm to patients. The underlying ques-tions are whether the law’s scope and applicationare appropriate, and if not, whether regulationof diagnostic tests and other medical devices canbe addressed differentially under the present lawor whether new legislative remedies should be ex-plored.

Second, it should be remembered that the find-ings that devices in higher regulatory classes havefewer manufacturers per device type and thatlarger manufacturers are more likely to manufac-ture devices in a higher regulatory class representthe situation that was already present before theamendments were implemented. Classificationunder the 1976 amendments did not cause butmight be expected to reinforce this situation, espe-cially for Class III devices, because of the highercosts associated with approval.

FDA also commissioned a survey conducted inthe fall of 1981 of medical device manufacturingestablishments that had been registered with FDAin September 1980 (197). The surveyors concludedthat there was no evidence that the amendmentsraised barriers to market entry, reduced innova-tion, or adversely affected investment, sales, oremployment. For example, one of the survey’sconclusions, based on information provided bythe surveyed manufacturers, was that there wasno evidence that patent activity had measurablydeclined since the Medical Device Amendmentswere enacted in 1976.

Figures 2 and 3 provide a more comprehensivepicture of medical device patent activities, sum-marizing patent applications with the U.S. Pat-ent Office between 1968 and 1979. Patent applica-tions on “low-technology” devices such as bandages,receptacles, eyeglass frames and lenses began lev-eling off just prior to the 1976 amendments. Butapplications for “high-technology” devices suchas implants, dialysis machines, respiratory de-vices, and cardiovascular devices continued to in-crease throughout the decade (see app. D)

The 1981 survey commissioned by FDA foundthat a third of all manufacturers had entered the

medical device field after the 1976 statute (197).Most manufacturers reported increases in domes-tic and foreign sales, research and development(R&D) activities, and the number of new devicesintroduced since the amendments. Fifty-one per-cent were more profitable and only 27 percent lessprofitable than they had been prior to passage ofthe amendments, and 80 percent were optimisticabout doing business in medical devices duringthe next decade.

The survey also found that significant R&Dactivities were common traits in medical devicefirms—whether they were large, medium, orsmall—and that the introduction of significantlynew medical devices had been just as common forsmall firms as for large firms (197). But when thesurvey was conducted in the fall of 1981, only aquarter of small establishments (1 to 9 employ-ees)—as compared to 63 percent of establishmentswith over 500 employees—reported that theywould consider developing and marketing a ClassIII device. The surveyors concluded that Class IIIdesignation appears to be more likely to discour-age small establishments than large establishmentsfrom developing new devices, but observed thatopinions do not necessarily translate into behav-ioral differences. They pointed out that 8.4 per-cent of establishments were manufacturing ClassIII devices and that a higher percentage of manu-facturers would continue developing Class IIIdevices.

Somewhat in contrast with the overall optimis-tic picture of the industry just presented weremanufacturers’ answers to the survey question ofthe impact of the Medical Device Amendments(197). Nearly half (46 percent) stated that Federalregulation had been a major problem for them,and 21 percent stated that regulation was thesingle most serious problem. However, althoughmost manufacturers wanted changes in the regu-lations, they did not believe (53 percent) that de-vice regulation should be abolished, and the vastmajority (80 percent) believed that at least im-plants and life-support or life-sustaining devicesshould be strictly regulated.

The specific problems associated with regula-tion under the 1976 amendments were varied(197). One problem reported by a substantial


Figure 2.–U.S. Patent Applications for Low-Technology Medical Devices, 1968-79

69 1970 172 1973 1974

Year of appl icat ion

1976I1977 1979

SOURCE: U.S. Department of Health and Human Services, Food and Drug Office of Economic Analysis, MD, compilation of unpublished data from the U.S. Patent and Trademark Office, December 1983.


Figure 3.–U.S. Patent Applications for High-Technology Medical Devices, 1968-79

U.S. origin

Foreign or ig in

1968 1969 1970 1971 1972 1973 1974

Year of appl icat ion

1978 1979

SOURCE: U.S. Department of Health and Human Services, Food and Administration, Office of Economic Analysis, MD, compilation of data from the U.S. Patent and Trademark Office, December 1983.


number of manufacturers in the 1981 survey wasthe cost of compliance. In order to meet the reg-ulatory requirements, 64 percent either added newemployees, purchased new equipment, or increasedoutside purchases. Absolute costs increased withestablishment size, but when adjusted for estab-lishment size, smaller manufacturers had rela-tively higher costs per employee in meeting theregulatory requirements.

Another problem reported by a significantnumber of manufacturers was in understandingwhat to expect from FDA in meeting the regula-tory requirements. Of particular interest is the cor-relation between manufacturers’ attitudes towardFDA and their understanding of the regulations.Of those manufacturers who said they fully un-derstood the regulations, about half (51 percent)gave FDA a positive rating, and 71 percent statedthat the regulations were effectively protecting thepublic. The Office of Small Manufacturers Assist-ance was one FDA information source that waspositively received by the industry, but difficultyin understanding the regulatory requirements wasstill a major problem and fell disproportionatelyon small manufacturers. Thus, a particular pri-ority for regulatory reform was in special effortsto improve manufacturers’ understanding of thedevice regulations.

Despite the negative opinions by manufacturersregarding regulation, majorities still reported thatregistration, product listing, product classifica-tion, labeling requirements, premarket approval,and IDEs had no effects on their establishments(197). Seventeen percent of manufacturers evenreported that good manufacturing practices havebeen of help to them.

Under the 1976 amendments, difficult and com-plex precedent-setting decisions have been madeon a diversified industry that was not previouslysubject to a great deal of FDA regulation. In gen-eral, the 1976 Medical Device Amendments havenot had a significant negative impact on the man-ufacturers of medical devices. Particular segmentsof the industry may be more affected than others,however, and compliance costs affect small man-ufacturers relatively more than they do large man-ufacturers.

In the contact lens industry, the issues of thecosts of complying with regulations and smallmanufacturers’ entry into the market have con-verged (see discussion under “Reclassification ofDevices” section, above). Class III designation ofnew types of contact lenses (soft lenses, gas-permeable lenses) has made it difficult for manysmall companies to gain early entry because ofthe costs of gathering clinical evidence on safetyand effectiveness. But there has not been a unifiedfront by the contact lens industry against Class111 designation. Rather, large firms that alreadyhave market approval have tended to resist reclas-sification from Class III to Class I or II. At issuein this instance is competition between first en-trants into the market and subsequent manufac-turers. The public policy goals that are at oddsare rewarding companies that first succeed in get-ting innovations on the market versus achievinggreater availability of products of a particulartype, with price competition as one result.

Throughout the medical devices industry, oneof the impacts of medical device regulation hasbeen uncertainty over the regulatory require-ments. This situation, in retrospect, is under-standable, given the fact that the implementation

Photo credit: Bausch & Lomb SOFLENS, Professional Products Division

The new generation of contact lenses, such as the soft con-tact lens show on the left, are subject to the full premarketapproval process of the Food and Drug Administration. Theolder types of hard contact lenses, such as that shown onthe right, no longer have to go through the full premarket ap-

proval process.

126 ● Federal Policies and the Medical Devices Industry—.-—- —

of some provisions of the law has not been initi-ated or completed and the fact that the majorityof devices have been placed in Class II, despiteinability to proceed with the statutory intent ofregulating this class of devices through perform-ance standards.

Conclusions

During the 8 years since the Medical DeviceAmendments of 1976 were enacted, the medicaldevices industry has continued to grow, and whileregulatory costs have been incurred, regulationhas generally not had a significant negative im-pact on the industry. A large part of the indus-try’s development may be due in part to FDA’simplementing the 1976 law in ways that wouldmake market entry easier—as in use of the 510kpremarket notification and a finding of “substan-tial equivalence” as the predominant route fordevices to be released for marketing—and in partto FDA’s not implementing or implementing slowlysome of the law’s provisions. The situation forindustry may change if FDA implements all of theprovisions of the amendments, or as new medi-cal devices are developed that make it harder overtime to use the “substantial equivalence” route tomarket devices.

Several provisions of the amendments that aretargeted at specific risk categories—such as thosepertaining to the safety and effectiveness of pre-amendments devices, regulation of Class II de-vices, and monitoring of devices once they are onthe market—have yet to be fully implemented oraddressed. Yet there is little information that ac-tual risks are systematically occurring or not be-ing addressed by FDA’s choice of priorities in im-plementing the amendments.

This paucity of information on actual risks canbe interpreted in two ways, based on opposing

POLICY OPTIONSMost of the attention that has been focused on

medical device regulation since the enactment ofthe 1976 Medical-Device Amendments has beenoriented toward questions such as whether a par-

assumptions. First, it might be taken as an indica-tion that hazards are in fact low, that the currentapplication of the amendments is satisfactory, andthat it is not necessary to implement all of thelaw’s provisions. An alternative interpretation isthat the paucity of information on risks is a defi-ciency in itself—one that the amendments attemptedto address—and that a lack of information onrisks is a problem that needs to be addressed.

The Medical Device Amendments providedmore effective methods for dealing with fraud-ulent devices, and the increasingly complex natureof “high-technology” medical devices was one ofthe imperatives for developing premarket screen-ing and testing requirements. Public policy inthese two instances was not primarily dependenton quantifying the number of injuries currentlycaused by medical devices. In the case of fraud-ulent devices, the amendments provided more ef-fective tools for removing these devices from themarket. For “high-technology” devices, theamendments attempted to anticipate and minimizepotential risks associated with their use throughpre- and postmarketing controls. Realistically,however, it might be expected that debates overhow and to what extent medical devices shouldcontinue to be regulated will focus on the coststo industry versus (lack of knowledge of) the ex-tent of risks associated with medical devices.

In sum, 8 years after the Medical Device Amend-ments of 1976 were enacted, the medical deviceindustry has incurred regulatory costs but con-tinues to prosper in general; major sections of thelaw remain partially or not implemented, andthere do not seem to be any obvious, major risksthat are not being addressed, a situation that mayreflect either a lack of significant risks or lack ofknowledge of significant risks that do exist.

ticular provision of the 1976 law has been imple-mented, whether its implementation has beencompatible with congressional intent, and whetherthe provision worked in practice as it did in con-


cept (331,338). A range of options proposed forspecific issues that the current law was designedto address is provided below. Areas of the lawto be specifically addressed include:

• evaluating the safety and effectiveness ofpreamendments devices and their postamend-ments equivalents,

● developing performance standards for ClassII devices,

. reviewing postmarketing activities and con-trols, and

. assisting small manufacturers of devices.

Beyond developing options on specific provi-sions of the law, however, there is the questionof how specific actions fit within an overall reg-ulatory framework. Various overall regulatory

approaches are presented in the first three optionsbelow.

Scope of Medical Device Regulation

Option 1: Continue the basic framework and in-tent of the 1976 Medical Device Amendmentsand make adjustments in implementation orwording of the specific provisions of the law.

A judgment could be made that the basic frame-work and intent of the 1976 amendments remainsappropriate and that the law’s implementation byFDA should proceed, subject to modifications inthe wording or implementation of specific provi-sions of the law that reflect judgments on theappropriate balance between methods of ensur-ing safety and effectiveness and the costs associ-ated with these methods.

FDA, in implementing the 1976 law, has hadto develop a set of priorities so that its limitedresources could be efficiently applied. Congresscould provide more direction to FDA on what itconsiders priority issues and what orientation itconsiders appropriate within a particular regula-tory area. Setting such priorities would entailweighing benefits to consumers from reducingrisks and ensuring efficacy versus costs to the in-dustry from regulatory requirements. Examplesof priority areas might include approval of newdevices, particularly Class III; safety and effec-tiveness of preamendments Class III devices; selec-tive monitoring and controls over marketed de-

vices; and development of better information ondevice-associated risks.

Congress could also provide direction withineach priority area on the extent of its concernabout ensuring safety and effectiveness versusminimizing barriers to market entry. Approachesto balancing safety and effectiveness versus easeof marketing are reflected by the variation in thetypes of safety and effectiveness evidence thatcould be required for preamendments Class IIIdevices and in whether FDA or device manufac-turers should bear the burden of proof (see Op-tions 4, 5, and 6).

A different strategy from focusing on whichprovisions of the law should be emphasized wouldbe for Congress to determine which aspects of thecurrent law do not have high priority. The ex-emption of devices from some of the law’s re-quirements through the use of FDA’s discretionaryauthority has been previously discussed. FDA hasexempted manufacturers of some Class I devicetypes (e. g., specimen containers) from having to

notify FDA when they intend to market theirdevices and from most of the recordkeeping re-quirements of the good manufacturing practicesregulations. In the regulations on IDEs, FDAmakes a distinction between “significant risk” and“nonsignificant risk” Class 111 devices and requiresdifferent procedures for the two. Also mentionedearlier was a petition to FDA from the ScientificApparatus Makers Association, subsequentlydenied by FDA as being against legislative intent,to drop 510k notification requirements for allClass I devices.

Option 2: Revise the 1976 Medical Device Amend-ments to reflect the status quo with regard toFDA’s implementation of the law.

Although the issuance of mandatory perform-ance standards for Class II devices has proved notto be feasible and FDA has yet to complete theimplementation of several other provisions of the1976 law, obvious, systematic deficiencies in thesafety and effectiveness of medical devices havenot been apparent. One approach, therefore,might be to recognize the two-tiered regulatoryapproach that has been applied to medical devicesrather than the three-tiered approach originallybuilt into the law.


More flexibility could be obtained through thekinds of controls identified previously to augmentor replace Class II performance standards (see Op-tion 13), but the bedrock of the law could belimited to: 1) general controls for all devices, and2) premarket approval requirements for a limitednumber of devices, such as implantable or life-supporting devices. Other current provisionscould also be modified or deleted. For example,review of preamendments devices could be limitedto high-priority device types, the approach thatFDA is currently taking.

Option 3: Revise the 1976 Medical Device Amend-ments to exclude certain device types from reg-ulation.

In the previous option, revisions in the lawwould be guided by FDA’s implementation of thelaw to date. In addition to or in place of that op-tion, Congress might choose to consider statutoryexclusions of some device types.

Statutory modifications could be guided byfocusing on risks, such as the proposal to exemptClass I devices from notification and recordkeep-ing requirements, or by focusing on the varietyof medical products currently under the jurisdic-tion of the amendments, such as the question ofwhether it is appropriate to regulate diagnostictests in the same manner as other types of medi-cal devices.

Regulation of Preamendments Devicesand Their Postamendments Equivalents

The 1976 amendments provided a 30-monthgrace period after final classification before evi-dence of the safety and effectiveness of preamend-ments Class III devices would be required by FDA.For these devices, two issues that remain are whattype of evidence has to be presented and whenthat evidence has to be provided to FDA. In part,these issues are important because of the wide-spread use of the “substantial equivalence” methodof gaining market entry for postamendmentsdevices. As previously discussed, a finding ofsubstantial equivalence will be made if a new de-vice does not differ markedly as to materials,design, or energy source, and if there is no sig-nificant difference with regard to safety and ef-

fectiveness. As yet, however, there is no require-ment to provide safety and effectiveness evidenceon the preamendments devices with which newdevices claimed to be their substantial equivalentsare compared.

In addition, FDA’s Office of General Counseldoes not consider a finding of “substantial equi-valence” an approval. A device is considered ap-proved once a determination is made that it is safeand effective. The 510k method of obtainingFDA’s permission to market a device is basicallya determination that the device is substantiallyequivalent to a preamendments device, and FDAhas no choice but to allow it to be marketed; itIS not a determination that the device is safe andeffective (464).

ISSUE:What evidence of safety and effectivenessshould be required of preamendmentsClass III devices?

Option 4: Continue FDA’s current approach ofemphasizing safety and effectiveness evidencefor high-priority preamendments Class IIIdevices.

Under FDA’s current policy as represented bythis option, preamendments Class III device typeswith questionable safety and effectiveness or withrelatively high risks will be addressed by FDAfirst, using expert opinion and publicly availableliterature. This approach can be viewed as a rea-sonable allocation of FDA’s limited resources, al-though FDA has to gather and review informa-tion to set areas of priority, and developing theinformation can be very resource-intensive forFDA

Option 5: Limit through legislation requirementsfor evidence of safety and effectiveness of pre-amendments Class III devices to device typesthat have specific problems associated withthem.

This option would codify FDA’s current ap-proach so that FDA would have to identify pre-amendments Class III device types with problemsbefore it could require evidence of safety and ef-fectiveness. Other preamendments device types

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration • 129— — . . —

would be presumed to be safe and effective, sub-ject to development of new information. As in theprevious option, this approach could be resource-intensive for FDA because the agency would haveto gather evidence to identify problem devices.Legislating this approach instead of relying onFDA’s discretion would reduce uncertainty andmake it explicit that all preamendments Class IIIdevices will not eventually have to show evidenceof safety and effectiveness.

Option 6: Encourage FDA to accept evidence ofsafety and effectiveness such as reviews of theliterature and expert opinion, in lieu of clini-cal evidence, for preamendments Class III devices.

In the two previous options, the safety and ef-fectiveness of preamendments Class III deviceswould in effect be presumed, and FDA would de-velop information to counter that presumptionbefore initiating actions. In this option, the bur-den of providing FDA with evidence of safety andeffectiveness would continue as now to rest withthe manufacturers, but the range of acceptabletypes of evidence would be greater. This approachwould enable FDA to screen all device types ora greater number than would the two previousoptions, and the screening process might then beused by FDA to target problem devices.

A variation of this option would be for FDAto start with the presumption that clinical dataon devices are required but allow manufacturersto overcome that presumption with evidencegained from general use of these types of devices.

ISSUE:When should safety and effectivenessevidence be required of preamendmentsClass 111 devices?

Option 7: Continue FDA’s interpretation that theend of the 30-month grace period after finalclassification establishes the earliest date thatFDA can require safety and effectiveness evi-dence on preamendments Class III devices.

Because the 30-month grace period establishesthe earliest date on which the agency can act, FDAhas begun the process of requiring safety and ef-fectiveness evidence for only a few “high-priority”preamendments Class 111 devices. FDA’s priority-

based review is dictated by the limited resourcesavailable to FDA and the resulting difficulty incalling for evidence of safety and effectiveness forall preamendments Class 111 devices as their graceperiods expire. Thus, the issue of when suchevidence will be required is related to the ques-tion of what kinds of evidence will be acceptable(see Options 4, 5, and 6).

Option 8: Establish the end of the 3&month graceperiod after final classification as the time whenFDA has to call for safety and effectivenessevidence on preamendments Class III devices.

This option could be legislated, but its desir-ability depends on whether FDA takes other ap-proaches to ensuring safety and effectiveness asdiscussed above and on the resources FDA coulddevote to preamendments devices relative to otherprovisions of the law. For example, if FDA takesthe approach in Option 6 of accepting a greaterrange of evidence to screen for problem devices,this option would be much more reasonable toimplement than under current conditions, inwhich FDA has assumed responsibility for iden-tifying problem areas.

Option 9: Prohibit use of the IDE to extend thegrace period for preamendments Class III de-vices that have been required to show evidenceof safety and effectiveness, except when noacceptable alternatives are available.

The grace period for many preamendmentsdevices had not ended or had not even begun asof early 1984, 8 years after the amendments werepassed. Given this extended period of “notifica-tion, ” there seems little justification for makingIDE routinely available to preamendments devicemanufacturers. Possibly, however, IDEs could bemade available on a case-by-case basis. Routineuse of the IDE to continue limited distribution ofpreamendments devices would be less of an issueif other types of evidence of safety and effective-ness, such as literature reviews and expert opin-ions, were accepted.

Except for those options specifically calling forlegislation, all of the options pertaining to pre-amendments Class 111 devices could be imple-mented under the existing statute. However, Con-gress could mandate a particular approach throughlegislative changes.

130 • Federal Policies and the Medical Devices industry. — —..——

ISSUE:Does the “substantial equivalence”method of entering the market forpostamendments medical devices needto be revised?

The 1976 Medical Device Amendments requirethat any postamendments device not found “sub-stantially equivalent” to a preamendments devicebe automatically classified in Class III, with subse-quent opportunity to petition for reclassificationof the device in Class I or II. The “substantialequivalence” clause of the 1976 law was meantto make a regulatory distinction between thosepostamendments devices that are modificationsof commercially available devices from those thatare truly new devices.

Because of the costs and delays in approvalassociated with the reclassification process, how-ever, manufacturers of postamendments deviceshave had incentives to seek a finding of “substan-tial equivalence” rather than reclassification sothat they can market their devices much sooner.Much less information is needed to successfullyclaim that a postamendments device is substan-tially equivalent to a preamendments device thanto gain approval through the premarket approvalprocess. In fact, the lack of information on thesafety and effectiveness of preamendments devicesraises questions about how determinations ofsubstantial equivalence can be made.

Option 10: Retain existing procedures for deter-mining ‘substantial equivalence. ”

As previously explained, FDA has begun to callfor safety and effectiveness evidence on high-priority preamendments Class III devices, andonce that evidence is presented and evaluated,there should be a substantive basis for compar-ing these devices with postamendments devicesdetermined to be substantially equivalent. Butthe process will take years and may be selectiverather than including all preamendments Class IIIdevices.

On the other hand, the “substantial equiva-lence” clause has been a convenient method fordevice manufacturers to get their products ontothe market quickly But as new generations of

postamendments devices diverge more and morefrom their preamendments antecedents, it will beharder for manufacturers to use the substantialequivalence method of market entry. It will alsobe harder to practice “piggybacking,” in whicha postamendments device is compared to anotherpostamendments device and, through a chain ofother postamendments devices, eventually com-pared to a preamendments device.

More immediately, FDA’s Office of GeneralCounsel has stated that such “piggybacking” is notauthorized by the amendments (464), and if thepractice of piggybacking ceases, more postamend-ments devices will eventually be placed in ClassIII, and their manufacturers will have to go throughthe full premarket approval process or petitionFDA for reclassification.

Option 11: Eliminate automatic classification intoClass III of postamendments devices that arenot found substantially equivalent to preamend-ments devices, and allow FDA to place a de-vice in the appropriate class at the time ofnotification.

Automatic classification into Class 111 of post-amendments devices that are not found substan-tially equivalent to preamendments devices servesas a second screen in the regulation of post-1976devices. The first screen is a determination ofwhether or not a device is “substantially equiva-lent” to a preamendments device. The secondscreen, with automatic classification into Class 111,is a presumption that any device that is not sub-stantially equivalent needs full premarket ap-proval, unless the manufacturer successfully peti-tions FDA for reclassification in Class I or Class11 Under this option, the burden of responsibilityof coming forth with evidence that rebuts initialClass 111 designation could remain with devicemanufacturers, but manufacturers could be al-lowed to present this evidence for classificationat the time of notification. This change should re-duce current incentives to claim “substantialequivalence. ”

Option 12: Develop approaches for reviewing newdevices that are different from those for review-ing modifications of commercially available .devices.

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration ● 131——

Eliminating automatic Class 111 designation ofpostamendments devices that are not found sub-stantially equivalent to preamendments devicesmight serve to bring out the distinction betweenmodified and new devices that the “substantialequivalence” clause was originally meant to pro-vide. GAO has recommended another approach,eliminating the “substantial equivalence” clauseso that all Class 111 (but not Class I or II) post-amendments devices have to go through premar-ket approval. (Automatic Class III designation fornon-substantially equivalent postamendments de-vices could also be eliminated so that new devicesthat would more appropriately be put into ClassI or II would not have to go through the super-fluous step of reclassification. ) In general, then,the difference in approach could be between pre-and postamendments devices as originally in-tended, or between Class III pre- and postamend-ments devices, where the difference in regulatoryrequirements is most pronounced.

Regulation of Class II, or Intermediate,Devices

More than 1,000 of the 1,700 device types havebeen placed in Class II. The unanimous opinion,however, is that except for a small number of de-vice types, performance standards cannot be de-veloped in a timely fashion. Thus, if an inter-mediate class of regulation is still needed, per-formance standards will have to be replaced byother types of regulation between Class I and itsgood manufacturing practices requirements andClass 111 and its premarket approval requirements.

Option 13: Give FDA legislative authority to useavailable methods in addition to performancestandards to regulate Class II devices.

An obstacle to the use of methods for regulat-ing Class II devices other than performance stand-ards has been the question of whether or not thelaw requires the use of performance standards.GAO has in fact suggested that the law be revisedto give FDA the authority to make a device-by-device determination of when performance stand-ards are needed (331). Although the use of per-formance standards may not be mandatory, achange in the statute clearing up the ambiguitymight be useful in setting into motion substan-

tive efforts to use other approaches, instead ofcontinuing to focus attention on the unrealisticexpectation that so many performance standardscan be developed.

FDA has suggested using a combination ofvoluntary standards, user education, and otherexisting controls to regulate Class II devices (387).Previously identified controls include revokingany approval to market a device, banning the de-vice, or requiring repair, refund, or replacement,and the prescription and restricted device provi-sions. If the legality of using available approachesin place of performance standards is upheld or theuse of these remedies legislated, a three-tiered reg-ulatory system for medical devices can be put inplace. Rather than a Class II with mandatory con-trols, however, there would be specific devices(Class I or Class III) for which additional controlscould be stipulated (e.g., prescription or restricteddevices), and device-by-device determinations ofthe applicability of these other controls.

Option 14: Legislate an additional category ofClass II devices to be regulated through meth-ods other than performance standards.

The Subcommittee on Oversight and Investiga-tions of the House Committee on Energy andCommerce has suggested that performance stand-ards be retained for Class II and that a Class 11Abe formed on which greater controls (e.g., restric-tions under the restricted device clause, increasedmandatory device experience reporting, and adop-tion of performance specifications against whichthe device must be tested periodically) are imposed(338).

This option is similar in effect to the previousoption, the principal difference being that this op-tion involves legislating an explicit, additional cat-egory of Class II devices and retaining mandatoryperformance standards for some devices. Also,this option would leave less discretion to FDA indetermining which devices should be regulatedand how they should be regulated.

Option 15: Encourage FDA to reclassify mostClass II device types into Class I or III and tocontinue to develop performance standards forthe remaining Class II devices.

Rather than being regulated through perform-ance standards, medical devices receiving Class

132 . Federal Policies and the Medical Devices Industry

II designation are currently being regulated byFDA as though they were Class I devices. A par-tial approach to the problem might be to screencurrent Class II devices to see whether some ofthem could be placed in Class III. This issue wasraised by the Health Research Group in the caseof General and Plastic Surgery devices, when FDAproposed placing seven implantable device types(including artificial ears, noses, and chins) in ClassII rather than Class III (253).

The burden of regulation under Class III forreclassified Class II devices might not be onerous.FDA already differentiates between “significantrisk” and “nonsignificant risk” Class III devicesin its IDE regulations; and FDA could develop dif-ferent levels of evidence for safety and effective-ness of the types previously discussed under op-tions for preamendments Class III devices (Option6).

Postmarketing Monitoring and Controls

The lack of information on risks associated withthe use of medical devices can be viewed eitheras evidence that such risks are not extensive andthat more vigorous device regulation is not needed,or instead as an indication that monitoring sys-tems should be improved to yield more informa-tion before risks are discounted. Identifying prob-lems is crucial in determining which devices mayneed additional controls and what types of con-trols should be applied. Thus, improved informa-tion on risks would be helpful both for determin-ing the scope of the problems that regulation couldaddress and in applying the appropriate types ofcontrols.

Option 16: Require FDA to develop better sys-tems for monitoring and providing informationon risks associated with devices.

FDA is reportedly ready to make final its reg-ulations on mandatory device experience report-ing by manufacturers, subject to the Office ofManagement and Budget’s approval (257). GAOhas suggested that FDA’s voluntary reporting sys-tem, the Device Experience Network (DEN), berevised so that information is included on thescope and nature of device problems caused byuser error and inadequate maintenance; that thedata be analyzed to identify special problems and

trends; and that the results be used to aid in de-veloping solutions (331). GAO is also initiatinga comprehensive exploration of postmarketingsurveillance activities and their potential applica-tions (300).

Thus, there is a gradual movement toward bet-ter identification of, and faster and more targetedresponses to, device risks. The process might beaccelerated by legislating mandatory device ex-perience reporting instead of continuing with thepermissive language contained in the statute.

Option 17: Encourage FDA to selectively applypostmarketing controls to regulate Class IIdevices.

Postarnendments controls could be applied toa new class of Class II devices or left to be ap-plied by FDA on a device-by-device basis (see Op-tions 13, 14, and 15). A reconstituted three-tieredclassification approach would result. Minimal reg-ulation would apply to the lowest class of devicesthrough the good manufacturing practices regu-lations. An intermediate class of devices (ClassII) would be represented by those devices thathave additional controls (prescriptions, restricteddevices, postmarked controls) applied to them ofthe types identified in addition to the good man-ufacturing practices requirements. The highestregulated class of devices would have to meetpremarket approval requirements and might haveadditional controls imposed on their marketing.

Assistance to Small Manufacturers

The 1976 amendments contained a provisionto help small firms through the regulatory proc-ess by establishing an Office of Small Manufac-turers Assistance. Two other steps could aid smallfirms in manufacturing Class III devices: 1) whereappropriate, Class III devices could be down-classified as soon as possible; and 2) small firmscould be given assistance in developing the safetyand effectiveness evidence necessary for Class IIIdevice approval.

Option 18: Develop additional mechanisms tohelp small firms through the regulatory process.

Option 18A: Encourage FDA to use publiclyavailable information as soon as possible todown-classify Class III devices.

Ch. 5—Regulation of Medical Devices by the Food and Drug Administration ● 133— ..—. —

FDA could take the initiative in identifyingClass III devices of significant importance to pub-lic health and could monitor their use. Thus, pub-licly available information could be accumulatedat the earliest possible time and down-classifica-tion could be initiated.

Option 18B: Develop a “broker” mechanism be-tween small firms with promising devices andclinical investigators capable of performing thetests necessary to gather safety and effective-ness data in support of the premarket approvalapplication for Class III devices.

Although Option 18A might help small firmsgain approval for medical devices that are alreadyon the market, it would not help small firms thatwant to be among the first to have their devicesapproved for marketing.

There is some precedent for a broker function,although there might be questions of conflict-of-interest if FDA assumed the role. One precedent,for example, is the past and continuing collabora-tion between commercial sponsors and specific in-stitutes at the National Institutes of Health in per-forming clinical trials for potentially significantnew drugs to meet FDA’s requirements of clini-cal testing for premarket approval. Another prece-dent is the Federal promotion of “orphan” medi-cal products, in which Federal funds are used tosupport clinical trials for promising products thathave a limited market, such as drugs for rare dis-eases. Thus, as a broker, FDA could maintain aregistry of potentially marketable devices and pro-vide it to interested parties.

regulation of medical devices by the food and drug administration

Documents